Kordin Neo
Organization Acme Labs
Search or jump to ⌘K
$21.05
A bell with a count and a short list of what needs attention is not in this build yet. Documentation Get help Signed in as Sam Okonkwo Your settings
Flow A. Hover a line to see its variable. Scale welcome

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout Degraded ↗ 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
api
Application, built from source · Ready
Public URL
https://k7f3q9x2.apps.kordin.cloud Open ↗
Inside the project
http://api.production:8080

Reachable only by the other services in this project. the public URL is real; the in-project name is a naming decision not yet taken

Where it comes from
Acme Labs/api
Branch main, deploying on every push. Built with Railpack · Node.js 24.
State
Ready
Answering on every instance.
rolled back · your next push to main deploys again
Deploys of checkout
a4f19c2
main · a moment ago
 Failed
3c9f0aa
main · serving now
 Ready
7be0d31
main · yesterday
1c88ae4
main · 3 days ago
What it reads
DATABASE_URL published by orders-db
Set here
LOG_LEVEL info
STRIPE_SECRET_KEY ••••••
Connected
orders-db
Postgres 16 · your own, reached over the network you set up.
DATABASE_URL published by orders-db, read by checkout

Managed databases are not part of October — this is a reference to a datastore you run yourself.

Public
https://k7f3q9x2.apps.kordin.cloud

The platform hostname. It is the only one in October.

Custom domains

Pointing your own domain at this service is not in October.

Pointing your own domain here is planned, once ownership and certificates can be verified.
Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

Machine
Size
What this service is charged for, for as long as it is deployed.
1 vCPU · 2 GB illustrative
Instances
More copies answer more traffic. Staged, and applied with everything else.
Deploys
On every push to main
Turning this off leaves deploying to you.
Build
Detected from the repository. Override only if the detection is wrong.
Railpack · Node.js 24▼
Removing
Remove this service
Removing is staged like everything else, and it runs last when you apply.
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout Degraded ↗ 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
checkout Degraded
Application, built from source
Public URL
https://k7f3q9x2.apps.kordin.cloud Open ↗
checkout is Degraded. A newer deployment did not become healthy, so the previous one keeps serving.
Logs
Reference 394b0e7c7626Get help ›
Show the code DESIRED_REVISION_NOT_SERVING
Inside the project
http://checkout.production:8080

Reachable only by the other services in this project. the public URL is real; the in-project name is a naming decision not yet taken

Where it comes from
Acme Labs/shop
Branch main, deploying on every push. Built with Railpack · Node.js 24.
State
Degraded
The latest deploy failed; the previous one keeps serving.
rolled back · your next push to main deploys again
Deploys of checkout
a4f19c2
main · a moment ago
 Failed
3c9f0aa
main · serving now
 Ready
7be0d31
main · yesterday
1c88ae4
main · 3 days ago
What it reads
DATABASE_URL published by orders-db
Set here
LOG_LEVEL info
STRIPE_SECRET_KEY ••••••
Connected
orders-db
Postgres 16 · your own, reached over the network you set up.
DATABASE_URL published by orders-db, read by checkout

Managed databases are not part of October — this is a reference to a datastore you run yourself.

Public
https://k7f3q9x2.apps.kordin.cloud

The platform hostname. It is the only one in October.

Custom domains

Pointing your own domain at this service is not in October.

Pointing your own domain here is planned, once ownership and certificates can be verified.
Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

Machine
Size
What this service is charged for, for as long as it is deployed.
1 vCPU · 2 GB illustrative
Instances
More copies answer more traffic. Staged, and applied with everything else.
Deploys
On every push to main
Turning this off leaves deploying to you.
Build
Detected from the repository. Override only if the detection is wrong.
Railpack · Node.js 24▼
Removing
Remove this service
Removing is staged like everything else, and it runs last when you apply.
◍ ▤ List

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main Replaced one at a time
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
1 change staged. 1 replaced one at a time
checkout — back to 7be0d31 Replaced one at a time Staged just now
◍ ▤ List

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout 1 instance · 512 MB Rolled back · your next push to main deploys again repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
◍ ▤ List

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
The rollback did not start. Nothing changed.
checkout cannot go back to 7be0d31 A secret it used, STRIPE_SECRET_KEY version 2, has been revoked. We will not quietly use a newer one: that would not be the deployment you asked for. Nothing changed, and 3c9f0aa is still serving. Reference 1b97933b2b2fGet help ›
Show the code ROLLBACK_SNAPSHOT_UNAVAILABLE
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api Updating ↗ 20s · replacing one at a time repo Acme Labs/api · main
checkout Updating ↗ 12s · 1 instance → 2 repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
search Provisioning 35s repo Acme Labs/search · main
Applying 3 changes. You can keep working; anything you add now waits for the next Apply.
◍ ▤ List
search Provisioning
Application, built from source · Provisioning
First deploy · 1m 12s · usually about 2 minutes
✓ Fetched your code ✓ Worked out how to build it ◐ Installing dependencies ○ Building ○ Starting it up
Following
14:02:11.204── Fetching Acme Labs/search at main@5d2e9b1
14:02:12.118Fetched 412 files (1.1 MB) in 0.6s
14:02:12.881── Working out how to build it
14:02:13.140Node.js 24.8.0 selected from engines field
14:02:14.006── Installing dependencies
14:02:18.774npm resolved 88 packages
Scheduling
Evaluating eligible capacity. 4s
Waiting for capacity
A valid request with no eligible capacity yet. You can cancel or change it. 3m 10s
Configuring
The network, the name and the configuration are converging. 1m 40s
Verifying
Health checks are running. 1m 58s
Cancelling
Cancelling safely; anything already made is being undone. 6s
Public URL, once it is live
https://q2m8v4r1.apps.kordin.cloud Open ↗
Where it comes from
Acme Labs/search
Branch main, deploying on every push.
rolled back · your next push to main deploys again
Deploys of checkout
a4f19c2
main · a moment ago
 Failed
3c9f0aa
main · serving now
 Ready
7be0d31
main · yesterday
1c88ae4
main · 3 days ago
What it reads
DATABASE_URL published by orders-db
Set here
LOG_LEVEL info
STRIPE_SECRET_KEY ••••••
Connected
orders-db
Postgres 16 · your own, reached over the network you set up.
DATABASE_URL published by orders-db, read by checkout

Managed databases are not part of October — this is a reference to a datastore you run yourself.

Public
https://k7f3q9x2.apps.kordin.cloud

The platform hostname. It is the only one in October.

Custom domains

Pointing your own domain at this service is not in October.

Pointing your own domain here is planned, once ownership and certificates can be verified.
Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

Machine
Size
What this service is charged for, for as long as it is deployed.
1 vCPU · 2 GB illustrative
Instances
More copies answer more traffic. Staged, and applied with everything else.
Deploys
On every push to main
Turning this off leaves deploying to you.
Build
Detected from the repository. Override only if the detection is wrong.
Railpack · Node.js 24▼
Removing
Remove this service
Removing is staged like everything else, and it runs last when you apply.

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
◍ ▤ List
llama-3.3-70b
Model, from the catalog · Ready
URL
https://m4t8x2p6.apps.kordin.cloud/v1
Call it from your code

curl https://m4t8x2p6.apps.kordin.cloud/v1/chat/completions -H "Authorization: Bearer $KEY" -d '{"messages":[{"role":"user","content":"Hello"}]}' the request shape is an example, not a published contract

Any API key with access to this environment works. API keys
Starting · 2m 40s
✓ Placed on 1×H100 ◐ Loading the model ○ Answering
Notify me when it answers October sends no notifications (11:439, V1-018)
you can fix this

Blocked. It does not fit on this environment's 1×H100 beside what is already on it. Remove another model from that GPU and it goes on.

Reference b51e70c2d8a4Get help ›
Show the code ACCELERATOR_SHARE_SHORTFALL
Runs on
1×H100
This environment's GPU. Only your own models share it.
Model
Llama 3.3 70B
From the catalog · Meta · 128k context

trying a model from its panel is Hasnat's ruling, not yet in forge

Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

Stop it
A stopped model gives its GPU back and keeps its URL. Staged like every change.

A model runs until you stop it. At $2.40 an hour this spends your $21.05 in about 9 hours, whether or not anything calls it. every rate is a shape, not a published price

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
◍ ▤ List
orders-db
Postgres 16 · managed

managed databases come later (V1-012, owner product owner); drawn so its place is kept

Storage
3.1 GB of 10 GB
Grows on its own; you can raise the ceiling, not lower it.
What it publishes
DATABASE_URL read by api and checkout
Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
◍ ▤ List
web Updating
Application, built from source · Updating
Public URL
https://w9c1d7e3.apps.kordin.cloud Open ↗
Deploying · 40s
✓ Built 3c9f0aa ◐ Replacing one at a time

A newer push to main is waiting and deploys after this one.

Inside the project
http://web.production:8080

Reachable only by the other services in this project. the in-project name is a naming decision not yet taken

Where it comes from
Acme Labs/web
Branch main, deploying on every push.
rolled back · your next push to main deploys again
Deploys of checkout
a4f19c2
main · a moment ago
 Failed
3c9f0aa
main · serving now
 Ready
7be0d31
main · yesterday
1c88ae4
main · 3 days ago
What it reads
DATABASE_URL published by orders-db
Set here
LOG_LEVEL info
STRIPE_SECRET_KEY ••••••
Connected
orders-db
Postgres 16 · your own, reached over the network you set up.
DATABASE_URL published by orders-db, read by checkout

Managed databases are not part of October — this is a reference to a datastore you run yourself.

Public
https://k7f3q9x2.apps.kordin.cloud

The platform hostname. It is the only one in October.

Custom domains

Pointing your own domain at this service is not in October.

Pointing your own domain here is planned, once ownership and certificates can be verified.
Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

Machine
Size
What this service is charged for, for as long as it is deployed.
1 vCPU · 2 GB illustrative
Instances
More copies answer more traffic. Staged, and applied with everything else.
Deploys
On every push to main
Turning this off leaves deploying to you.
Build
Detected from the repository. Override only if the detection is wrong.
Railpack · Node.js 24▼
Removing
Remove this service
Removing is staged like everything else, and it runs last when you apply.

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
◍ ▤ List
mailer
Application, from an image · Ready
Inside the project
http://mailer.production:8080

Reachable only by the other services in this project. the in-project name is a naming decision not yet taken

Where it comes from
ghcr.io/Acme Labs/mailer:2.1
An image, pulled when it deploys. Listens on port 8080.
State
Ready
Answering on its one instance.
rolled back · your next push to main deploys again
Deploys of checkout
a4f19c2
main · a moment ago
 Failed
3c9f0aa
main · serving now
 Ready
7be0d31
main · yesterday
1c88ae4
main · 3 days ago
What it reads
DATABASE_URL published by orders-db
Set here
LOG_LEVEL info
STRIPE_SECRET_KEY ••••••
Connected
orders-db
Postgres 16 · your own, reached over the network you set up.
DATABASE_URL published by orders-db, read by checkout

Managed databases are not part of October — this is a reference to a datastore you run yourself.

Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

Machine
Size
What this service is charged for, for as long as it is deployed.
1 vCPU · 2 GB illustrative
Instances
More copies answer more traffic. Staged, and applied with everything else.
Deploys
On every push to main
Turning this off leaves deploying to you.
Build
Detected from the repository. Override only if the detection is wrong.
Railpack · Node.js 24▼
Removing
Remove this service
Removing is staged like everything else, and it runs last when you apply.

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
◍ ▤ List
sessions
Redis 7 · managed

managed databases come later (V1-012, owner product owner); drawn so its place is kept

Storage
240 MB
Grows on its own; you can raise the ceiling, not lower it.
What it publishes
REDIS_URL read by web
Last hour
Requests
240 a minute · p95 84 ms
Memory
318 MB

nothing reports these yet, and what a metrics pane holds is a product decision, not a backend one

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout ↗ 1 instance · 512 MB repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
api Ready
Where it comes from
Acme Labs/api · branch main, deploying on every push
What it reads
DATABASE_URL from orders-db
Machine
2 instances · 512 MB
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main Replaced one at a time
checkout ↗ 1 instance → 2 · 512 MB repo Acme Labs/shop · main Can’t know yet
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
search New Nothing built yet repo Acme Labs/search · main Nothing running is touched
3 changes staged. 1 replaced one at a time · 1 touches nothing running · 1 can’t know yet
search — added from Acme Labs/search · main Nothing running is touched Staged 2 minutes ago
api — STRIPE_SECRET_KEY value replaced · version 3 → 4 Replaced one at a time Staged 2 minutes ago
Was
value replaced · version 3
Now
value replaced · version 4
checkout — 1 → 2 instances Can’t know yet Staged 6 minutes ago Changed since staged — a new deployment of checkout landed 3 minutes ago. Restage on top of it, or drop this row.
a note on a change comes later (22 D4)
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api Updating ↗ 20s · replacing one at a time repo Acme Labs/api · main
checkout Updating ↗ 12s · 1 instance → 2 repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
search Provisioning 35s repo Acme Labs/search · main
Applying 3 changes. You can keep working; anything you add now waits for the next Apply.
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout ↗ 2 instances · 512 MB repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
search 1 instance · 512 MB repo Acme Labs/search · main
3 changes applied. search is live k7f3q9x2.apps.kordin.cloud Open ↗ 1m 52s from Apply to live · Every push to main deploys on its own. apiKey updated; replaced one at a time. checkoutNow 2 instances.
checkout did not roll out.

The new revision did not become healthy before its progress deadline. The rollout paused or rolled back automatically and the previous revision keeps serving.

Show the code ROLLOUT_PROGRESS_DEADLINE_EXCEEDED
2 of 3 applied. checkout did not roll out.

The new revision did not become healthy before its progress deadline. The previous revision (one instance) keeps serving.

Show the code ROLLOUT_PROGRESS_DEADLINE_EXCEEDED
search is live apiKey updated; replaced one at a time.
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout ↗ 1 instance → 2 · 512 MB repo Acme Labs/shop · main Can’t know yet
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
search 1 instance · 512 MB repo Acme Labs/search · main
2 of 3 applied. search and api are live. checkout’s change is waiting for you.
checkout changed since you staged this — 3 minutes ago You staged 1 → 2 instances. Now: a new deployment, 1 instance.
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout Degraded ↗ 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
production
Environment
General
What runs here2 services
All services▼
checkoutDegraded — previous deployment serving Degraded
apiReady Ready
HardwareNone dedicated

Nothing dedicated in this environment yet.

Who can reach itPublic URLs
Every service's public URL
Each service answers on its own platform hostname. Nothing here is private, and there is no per-environment network setting to change in October.
LogsAll services
All services▼

checkout · the deployment paused, the previous one kept serving
api · listening on :8080
web · listening on :8080

Variables2 set above the service
Variables set above the service
Key Set at
LOG_LEVEL environment
DATABASE_URL orders-db (published)
Danger
Remove this environment
Removing an environment is staged like everything else, and it runs last when you apply.
1 change staged. Nothing has happened yet. Applying grows sessions from 240 MB to 1 GB.
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main Replaced one at a time
checkout ↗ 1 instance → 2 · 512 MB repo Acme Labs/shop · main Can’t know yet
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
search New Nothing built yet repo Acme Labs/search · main Nothing running is touched
3 changes staged. Nothing has happened yet and nothing is charged until they land. Applying adds search, restarts api, and takes checkout from one instance to two.
◍ ▤ List
production
Environment
General
What runs here2 services
All services▼
checkoutDegraded — previous deployment serving Degraded
apiReady Ready
HardwareNone dedicated

Nothing dedicated in this environment yet.

Who can reach itPublic URLs
Every service's public URL
Each service answers on its own platform hostname. Nothing here is private, and there is no per-environment network setting to change in October.
LogsAll services
All services▼

checkout · the deployment paused, the previous one kept serving
api · listening on :8080
web · listening on :8080

Variables2 set above the service
Variables set above the service
Key Set at
LOG_LEVEL environment
DATABASE_URL orders-db (published)
Danger
Remove this environment
Removing an environment is staged like everything else, and it runs last when you apply.

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL REDIS_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main Replaced one at a time
checkout ↗ 1 instance → 2 · 512 MB repo Acme Labs/shop · main Can’t know yet
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
search New Nothing built yet repo Acme Labs/search · main Nothing running is touched
◍ ▤ List
production
List · 8 services
All kinds▾
Services in production
Service State Running Last deploy
Application · Staged · restarts when applied
 Ready 2 instances main@3c9f0aa
12m ago
Application · Staged · 1 instance → 2, does not restart
Degraded 1 instance main@9b1c4d2
33m ago
Application · Staged · added, not applied
 Not built yet — not yet
Application
 Ready 2 instances main@a01f7e2
3h ago
Application
 Ready 1 instance mailer:2.1
2 days ago
Model
 Ready 1×H100 —
Data
 Ready 3.1 GB of 10 GB —
Data
 Ready 240 MB —
3 changes staged. Nothing has happened yet and nothing is charged until they land. Applying adds search, restarts api, and takes checkout from one instance to two.

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
DATABASE_URL DATABASE_URL
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout ↗ 1 instance · 512 MB repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
DATABASE_URL
Published by orders-db
Read by api, checkout
Not read by web, mailer
◍ ▤ List

production

Acme Labs›acme-web›· 6 services $0.12 an hour 1×H100 dedicated
Add something
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
Open url ↗ web ↗ 2 instances · 512 MB repo Acme Labs/web · main
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
Add something Where it runs
◍ ▤ List

production

Acme Labs›acme-web›· 0 services $0.12 an hour 1×H100 dedicated
Add something
orders-db Your own data Postgres 16 · bring your own
sessions Your own data Redis 7 · bring your own
api ↗ 2 instances · 512 MB repo Acme Labs/api · main
checkout ↗ 1 instance · 512 MB repo Acme Labs/shop · main
web ↗ 2 instances · 512 MB repo Acme Labs/web · main
sessions Ready
Remove this data service
The data goes, the backups go, and REDIS_URL is removed from the environment — so web, which reads it, fails on its next start. That is the dotted line on the drawing.
Nothing is deleted yet. Removing here adds it to the staged changes, and it runs last when you apply, because it is the one step that cannot be undone.
Read by
web · one service. Unbind it there first and this becomes a change that costs nothing.
◍ ▤ List

production

Acme LabsScale›acme-web›1 drift· 6 services $0.12 an hour 1×H100 dedicated
Workloads you run with kubectl are not shown here.
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
production
Environment
KubernetesScale
Scale has ended for Acme Labs. What is here keeps running, and you can still read it and make it smaller. This feature is not included in the current entitlement. Ask the account owner to enable it.
Reference 394b0e7c7626Get help ›
Show the code CAPABILITY_NOT_GRANTED
Something here was changed by hand, and it conflicts with what we manage. Automatic management of that part is paused until it is put back.
Reference 394b0e7c7626Get help ›
Show the code forge names no reason code for a plain drift condition — for-forge.md #10; forge/docs/v1/12-runtime-tenancy-and-workloads.md:156 "Conflicting customer changes produce a visible drift condition"
No Kubernetes yet
Version The one version offered here. 1.31
Machines One size, standard-4-16 forge/contracts/product_classes.yaml:60 — the October node shape's key is EVIDENCE_REQUIRED; "October publishes one class"
Kubernetes Staged
Version 1.31
Machines 3 · standard-4-16

Waits in the strip with everything else. Nothing is built until you apply.

Kubernetes 1.31 Provisioning

The Kubernetes control plane for this environment is still being created. No action is needed; progress continues automatically.

Kubernetes 1.31 Ready
Worker pools
Direct access
Kubernetes 1.31 Ready

No worker pools, so nothing you run with kubectl has anywhere to go.

Kubernetes 1.31 Updating

Moving to a new set of machines. Work moves before the old ones go.

Kubernetes 1.31 Deleting

Going with the environment, which is being removed.

Kubernetes 1.31 Blocked
A Kubernetes control plane that was previously established is no longer present. Recreating it automatically could destroy state, so an operator reviews it before anything changes.
Reference 394b0e7c7626Get help ›
Show the code CONTROL_PLANE_LOST
What stays ours

Even kubernetes.owner cannot mutate Nodes, cluster-scoped RBAC, CRDs, admission policy, Cilium/Kordin system resources, platform Gateway/certificate objects or reserved ingress labels. Admission policy also denies privileged, host-network and hostPath pods and binding the reserved environment-Gateway port.

Read more in the docs ›

Worker pools in production
Pool Machines Running
general
standard-4-16
 3 32
Waiting for capacity. Part of the requested capacity is not available yet. What has already been created is kept while the remainder is retried in the background.
Show the code CAPACITY_SHORTFALL
Machines in general Staged, and applied with everything else.

No new access while Scale has ended. Anything already issued stops within 10 minutes.

Your access was taken away. No new access from now. Anything already issued stops within 10 minutes.
~/.kube/config 
# There is no key in this file.
clusters:
  - cluster:
      server: https://production-012312.k8s.kordin.cloud
      certificate-authority-data: LS0tLS1CRUdJTiBD…
users:
  - user:
      exec:
        command: kordin
        args: ["auth", "k8s-token", "--environment=production"]
Then run forge names no CLI commands (topic 11); the exec-plugin flow is forge/docs/v1/04-security-and-trust.md:236
kubectl get pods

Each command gets a token that lasts 5 to 10 minutes, for your role in this environment only, never admin.

General
What runs here2 services
All services▼
checkoutDegraded — previous deployment serving Degraded
apiReady Ready
HardwareNone dedicated

Nothing dedicated in this environment yet.

Who can reach itPublic URLs
Every service's public URL
Each service answers on its own platform hostname. Nothing here is private, and there is no per-environment network setting to change in October.
LogsAll services
All services▼

checkout · the deployment paused, the previous one kept serving
api · listening on :8080
web · listening on :8080

Variables2 set above the service
Variables set above the service
Key Set at
LOG_LEVEL environment
DATABASE_URL orders-db (published)
Danger
Remove this environment
Removing an environment is staged like everything else, and it runs last when you apply.
1 change staged. Kubernetes 1.31 with 3 machines · can’t know yet
◍ ▤ List

production

Acme LabsScale›acme-web›1 drift· 6 services $0.12 an hour 1×H100 dedicated
Workloads you run with kubectl are not shown here.
Add something
DATABASE_URL DATABASE_URL REDIS_URL
orders-db Your own data Postgres 16 · bring your own
orders-db
Postgres 16
Your own
Read by api, checkout
Click to open everything
sessions Your own data Redis 7 · bring your own
sessions
Redis 7
Your own
Read by web
Click to open everything
Open url ↗ api ↗ 2 instances · 512 MB repo Acme Labs/api · main
api
Ready · Acme Labs/api
2 instances · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ checkout Degraded 1 instance · 512 MB Did not become healthy — the previous deployment keeps serving repo Acme Labs/shop · main
checkout
Degraded · the previous deployment keeps serving
1 instance · 512 MB
Reads DATABASE_URL
Click to open everything
Open url ↗ web Updating 40s · 3c9f0aa · + a newer push waiting repo Acme Labs/web · main
web
Ready · Acme Labs/web
2 instances · 512 MB
Reads REDIS_URL
Click to open everything
Open url ↗ llama-3.3-70b ↗ 1×H100 · answering model Llama 3.3 70B
llama-3.3-70b
Answering
1×H100
Llama 3.3 70B
Click to open everything
Open url ↗ mailer Not reachable from outside image ghcr.io/Acme Labs/mailer:2.1
mailer
Ready
Not reachable from outside
ghcr.io/Acme Labs/mailer:2.1
Click to open everything
Add something Where it runs
You can add Kubernetes to this environment.
◍ ▤ List

production

Acme Labs›acme-web›
Add something

An empty environment

The drawing fills itself in as you add things. There is nothing to arrange and nothing to connect by hand — a line appears once one thing reads a variable another publishes.

production

Acme Labs›acme-web›
Add something

We could not draw this environment

The service could not complete the request right now. Retry with the same idempotency key after the interval given with this response.

This is a read: the shape did not come back.

Nothing was disturbed.

Reference 7f3a2c19e04bGet help ›
Show the code TEMPORARILY_UNAVAILABLE