Self-host PostHog

I have self hosted posthog on GCP and integrated into my next js application currently am facing "Forbidden (403) CSRF verification failed. Request aborted" for the below api calls where events are getting captured Request URL: https://{Url}/i/v0/e/?ip=1&_=1741776541825&ver=1.224.0&compression=gzip-js Request Method: POST Status Code: 403 Forbidden

Thanks for fixing the recent issues with cyclotron and clickhouse. I finally got the self-hosted app running and have signed up successfully. However, I can't get past the login screen. Looking at the web-1 logs, the login is successful (username and pw is okay) but eventually authorization fails at the requests to /api/billing/get_invoices and /api/billing. Did I miss a step? Any advice what I can do?

(Not sure if it has anything to do with it, I'm accessing the app running on the EC2 instance via http://{IP address}, not using https://{domain name}. I'm using Experimental mode.)

Unless you are using a quantum computer

I have self deployed and the event tracking is working fine but i can't search anywhere directly I have checked the network tab the search api give no result but HogQL work and give applies that filter what should I do in this case ?

The following error happens:

  File "/code/posthog/async_migrations/migrations/0002_events_sample_by.py", line 154, in is_required
    table_engine = sync_execute(
                   ^^^^^^^^^^^^^
IndexError: list index out of range

I checked the code, this line tries to query table "events" while it's not created in postgres. Then I checked an previous error, it seems the container is trying to send requests to clickhouse, but using the "localhost:9000" which seems to be wrong since localhost is pointing to container itself instead of clickhouse container.

    |   File "/python-runtime/clickhouse_driver/connection.py", line 417, in connect
    |     raise err
    | clickhouse_driver.errors.NetworkError: Code: 210. Connection refused (127.0.0.1:9000)

The only change I made to clickhouse container is changing it's tag to 24.11.3.66 because version 24.10 encountered an error

the web-1 container has logs like:

{'event': "\n                You indicated your instance is behind a proxy (IS_BEHIND_PROXY env var),\n                but you haven't configured any trusted proxies. See\n                https://posthog.com/docs/configuring-posthog/running-behind-proxy for details.\n            ", 'timestamp': '2025-02-07T16:48:52.541553Z', 'logger': 'posthog.settings.access', 'level': 'warning', 'pid': 7, 'tid': 140265971866496}

the container cyclotron-fetch-1 and cyclotron-janitor-1 starts failed. Docker logs shows that "database "cyclotron" does not exist"

should I manually create the database?

I've got errors of forbidden and CORS to my self-hosted PostHog already tried to add ALLOWED_HOST inside the docker-compose.yml for the web container but the issues are still not solved

Validation checks

9 successful, 1 error

Plugin server · Node is showing error.

any way to debug it? https://tinyurl.com/ymlzlfso

I tried setup of PostHog and it gave below error:

"Error response from daemon: Get "https://ghcr.io/v2/": dialing ghcr.io:443 with direct connection: resolving host ghcr.io: lookup ghcr.io: no such host"

And there was also Kakfa error.

Any solution to fix this please?

Running the set up command is failing with the following error:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/posthog/posthog/HEAD/bin/deploy-hobby)"

Grabbing latest apt caches The operation couldn’t be completed. Unable to locate a Java Runtime that supports apt. Please visit http://www.java.com for information on installing Java.

I recently installed the latest version of posthog. Now I am just seeing a spinning circle

I am seeing HTTP 500 returned for /api/billing-v2/. On checking the web logs it is giving the following error

File "/code/ee/billing/billing_manager.py", line 177, in get_billing if not addon["subscribed"]: KeyError: 'subscribed'

How can I fix this issue ? I am using the installation script mentioned in this doc

I am trying to upgrade posthog JS to latest version. I deleted all my existing containers and re-ran the installation script mentioned in this doc to bring posthog up again. The package json in the downloaded files in posthog/ folder is saying posthog-js version as 1.131.2 but I am still getting served version 1.118.0. Why is this happening ? How do I get the absolute latest version of posthog in all components?

I am trying to use an external clickhouse cluster with self hosted posthog. I have given my clickhouse hot in the environment variable. But i am not getting any tables created in my clickhouse server. The database posthog got created correctly. The cluster name is also posthog. If I keep the environment variables with my host value but also bring the built in clickhouse container up, I can see tables are created there. Is the clickhouse host url hardcoded somewhere in the code ?

I followed this article and got the containers up and running. Now my SSL wasnt working for some reason so I re ran the container with debug: 'true'. Now my frontend is not loading. I can see it is looking for JS files at mydomain.com:8234. It is failing with network error. I cant see any containers up in port 8234 in my machine. Any clue on this ?

We are currently experiencing an issue with the PostHog upgrade script for our self-hosted version. We have attempted to use the following script to upgrade PostHog:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/posthog/posthog/HEAD/bin/upgrade-hobby)"

However, during the execution of the script, we encountered the following error:

Error response from daemon: invalid reference format

It appears that the script suddenly stopped with this error message. We have followed the provided instructions for upgrading, but we seem to be encountering this obstacle.

Could you please guide us how we can successfully update our self-hosted version of PostHog? We appreciate any assistance or insights you can offer to resolve this issue promptly.

Error screenshot: https://img.enacton.com/ShareX/2024/02/putty_qIlxKd3LD1.png

Thank You.

Everything is fine until it sets up docker compose...

Error response from daemon: invalid reference format

Full output

Upgrading PostHog. This will cause a few minutes of downtime.
Do you want to upgarde PostHog? [y/N] y
OK!
Checking for named postgres and clickhouse volumes to avoid data loss when upgrading from < 1.39
Found postgres and clickhouse volumes, proceeding...
Already up to date.
Setting up Docker Compose
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
100 43.5M  100 43.5M    0     0   114M      0 --:--:-- --:--:-- --:--:--  114M
Error response from daemon: invalid reference format

Is it possible to have a detailed guide on how to set the secret key. IT IS NOT CLEAR

We are currently running a VM with nginx and I need to set up PostHog under our proxy. However I can’t find how to change the PostHog web port. The installation crashes because PostHog wants to access a running port.

So as the enterprise version is sunset aka discontinued for onprem, how can we create multiple projects on selfhosted version.

Currently on selfhosted version we can only create one project and organization.

We want to setup multiple websites in the instance but don't want to create multiple instances for the same.