Running Langtrace Client on Railway
Resource Type | Comments |
---|---|
Clickhouse Service | To store traces. |
Postgres Service | For login and dataset related info. |
Langtrace Client | Langtrace app with frontend. |
Setting up Langtrace on Railway is now effortless. Just follow these simple steps:
When deploying the template, you’ll need to set up authentication for your Langtrace instance:
langtrace-client
service.ADMIN_EMAIL
: Your desired admin emailADMIN_PASSWORD
: Your desired admin passwordlangtrace-client
service and add admin email and admin password and click “Save Config” → “Deploy”.
Once you start the deployment process, it may take around two minutes for Railway to bring up the services.
If you want you can check the deploy logs of a particular service by clicking on the service and click “View Logs” under the Deployments tab.
Once all services are healthy, click on ‘langtrace-client’ service and open the URL to access the self-hosted Langtrace application.
Login to Langtrace using the previously supplied Admin email and password, and begin your journey for monitoring LLM apps by following the quickstart guide.
To cleanup the resources, click on Settings —> Danger Zone —> Delete Project.
Clickhouse service keeps restarting or crashing
This generally occurs due to low memory allocated to the clickhouse service. You can increase the memory allocated to the clickhouse service. This may require you to upgrade your Railway plan from the free tier.
Update Langtrace version
Click on the langtrace-client
service —> you should be in Deployments tab —> Select Redeploy in the option of the active deployment to allow Railway to pull the latest version of the Docker image.
OR
Click on the langtrace-client
service —> go to Settings tab —>
disconnect —> connect image —> change source image to any version of your choice —> Click Deploy
For example: scale3labs/langtrace-client:1.3.1
Running Langtrace Client on Railway
Resource Type | Comments |
---|---|
Clickhouse Service | To store traces. |
Postgres Service | For login and dataset related info. |
Langtrace Client | Langtrace app with frontend. |
Setting up Langtrace on Railway is now effortless. Just follow these simple steps:
When deploying the template, you’ll need to set up authentication for your Langtrace instance:
langtrace-client
service.ADMIN_EMAIL
: Your desired admin emailADMIN_PASSWORD
: Your desired admin passwordlangtrace-client
service and add admin email and admin password and click “Save Config” → “Deploy”.
Once you start the deployment process, it may take around two minutes for Railway to bring up the services.
If you want you can check the deploy logs of a particular service by clicking on the service and click “View Logs” under the Deployments tab.
Once all services are healthy, click on ‘langtrace-client’ service and open the URL to access the self-hosted Langtrace application.
Login to Langtrace using the previously supplied Admin email and password, and begin your journey for monitoring LLM apps by following the quickstart guide.
To cleanup the resources, click on Settings —> Danger Zone —> Delete Project.
Clickhouse service keeps restarting or crashing
This generally occurs due to low memory allocated to the clickhouse service. You can increase the memory allocated to the clickhouse service. This may require you to upgrade your Railway plan from the free tier.
Update Langtrace version
Click on the langtrace-client
service —> you should be in Deployments tab —> Select Redeploy in the option of the active deployment to allow Railway to pull the latest version of the Docker image.
OR
Click on the langtrace-client
service —> go to Settings tab —>
disconnect —> connect image —> change source image to any version of your choice —> Click Deploy
For example: scale3labs/langtrace-client:1.3.1