Re: How to set Liveness & Readiness probe for pgadmin running in Kubernetes

Hi Aruna,

Next pgAdmin 4 release is on 4th May. As I suggested earlier you can try nightly builds to test the issue.

#6120 occurred due to changes done in #5567.

Thanks,

Yogesh Mahajan

EnterpriseDB

On Wed, Apr 19, 2023 at 7:24 PM Aruna Manivannan <aruna.manivannan@gmail.com> wrote:

Hi Yogesh, 

Thanks again. I can see that issue #5567 is closed. So for both issue #6120 &  #5567 the fix is the same ? 
Have provided the required screenshots in the thread . Can you please let me know when the next release date is? Thank you.

Regards,

Arun

On Wed, Apr 19, 2023 at 8:03 AM Yogesh Mahajan <yogesh.mahajan@enterprisedb.com> wrote:

Hi Aruna,

We have fixed #6120 issue where users were getting similar csf errors which were happening in case of authsource = oauth2.

Fix will be available in the next release. Meantime you want to verify the issue, you use a nightly build docker image. (docker pull dpage/pgadmin4:snapshot)

Thanks,

Yogesh Mahajan

EnterpriseDB

On Wed, Apr 19, 2023 at 4:52 PM Aruna Manivannan <aruna.manivannan@gmail.com> wrote:

Hi Yogesh, 

We don't have any settings in config_local.py as there is no such file and the authentication mode is SSO via okta. 

AUTHENTICATION_SOURCES = ['oauth2','internal'] 

Thanks. 

Regards,

Arun

On Wed, Apr 19, 2023 at 1:41 AM Yogesh Mahajan <yogesh.mahajan@enterprisedb.com> wrote:

Hi,

Could you please share config_local.py settings? Which authentication mode are you using?

Thanks,

Yogesh Mahajan

EnterpriseDB

On Tue, Apr 18, 2023 at 7:23 PM Aruna Manivannan <aruna.manivannan@gmail.com> wrote:

Hi Yogesh, 

Please find the attached config file which we are using . 
The issue which we are facing is exactly the same as it is described in this thread and my team member Eoghan also commented on the same. 

Below are the parameters that were modified according to our needs , rest were all the same default values only. 

 # Number of values to trust for X-Forwarded-For

PROXY_X_FOR_COUNT = 1

 # Number of values to trust for X-Forwarded-Proto.

PROXY_X_PROTO_COUNT = 1

 # Number of values to trust for X-Forwarded-Host.

PROXY_X_HOST_COUNT = 1

 # Number of values to trust for X-Forwarded-Port.

PROXY_X_PORT_COUNT = 0

 # Number of values to trust for X-Forwarded-Prefix.

PROXY_X_PREFIX_COUNT = 0

USER_INACTIVITY_TIMEOUT = 900

ENHANCED_COOKIE_PROTECTION = False

Regards,

Arun

On Tue, Apr 18, 2023 at 7:23 AM Dave Page <dpage@pgadmin.org> wrote:

On Tue, 18 Apr 2023 at 12:08, Yogesh Mahajan <yogesh.mahajan@enterprisedb.com> wrote:

Hi,

Could you please share your pgAdmin config setting?

FYI, the /misc/ping endpoint is CSRF-exempt.

 

Thanks,

Yogesh Mahajan

EnterpriseDB

On Fri, Apr 14, 2023 at 8:45 PM Aruna Manivannan <aruna.manivannan@gmail.com> wrote:

Hi Yogesh, 

Thanks for your reply. Please find the attached log file. The log entries keep repeating so given the portion of them. 

Sending it to you alone due to security reasons.
FYI we have the parameter ENHANCED_COOKIE_PROTECTION = false to avoid CSRF token issue, but it seems to be ignored. 
Thanks. 

Regards,

Arun

On Fri, Apr 14, 2023 at 5:33 AM Yogesh Mahajan <yogesh.mahajan@enterprisedb.com> wrote:

Hi,

Could you please complete the logs from the container? I suspect those errors are because of CSRF token errors.

Thanks,

Yogesh Mahajan

EnterpriseDB

On Thu, Apr 13, 2023 at 11:04 PM Aruna Manivannan <aruna.manivannan@gmail.com> wrote:

Hi Yogesh / pgadmin-support team, 

Have set up the liveness probe as suggested by you. However, It is now super helpful. The problem I am facing is though the /misc/ping is giving 200 status, I received complaints from users that they cant access pgadmin.
When I checked the logs, I noticed that /sqleditor/status/<some number> is 500 & 503. I guess the number could be session ID or something not sure. Please find the below log message 

[::ffff:<private_ip_address> - - [13/Apr/2023:16:28:13 +0000] "GET /misc/ping HTTP/1.1" 200 4 "-" "kube-probe/1.24+"

::ffff:<private_ip_address> - - [13/Apr/2023:16:28:13 +0000] "GET /misc/ping HTTP/1.1" 200 4 "-" "kube-probe/1.24+"
<public_ip_address> - - [13/Apr/2023:16:28:14 +0000] "GET /sqleditor/status/4397343 HTTP/1.1" 500 103 "https://pgadmin-url/sqleditor/panel/4397343?is_query_tool=true&sgid=96&sid=307&did=<db_id>&database_name=<db_name>" "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/112.0"
<public_ip_address> - - [13/Apr/2023:16:28:14 +0000] "GET /sqleditor/status/9414096 HTTP/1.1" 503 100 "https://pgadmin-url/sqleditor/panel/9414096?is_query_tool=true&sgid=96&sid=307&did=<db_id>&database_name=<db_name>" "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/112.0"
<public_ip_address> - - [13/Apr/2023:16:28:14 +0000] "GET /sqleditor/status/9414096 HTTP/1.1" 503 100 "https://pgadmin-url/sqleditor/panel/9414096?is_query_tool=true&sgid=96&sid=307&did=<db_id>&database_name=<db_name>" "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/112.0"
<public_ip_address> - - [13/Apr/2023:16:28:14 +0000] "GET /sqleditor/status/9414096 HTTP/1.1" 503 100 "https://pgadmin-url/sqleditor/panel/9414096?is_query_tool=true&sgid=96&sid=307&did=<db_id>&database_name=<db_name>" "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/112.0

Later I restarted the pgadmin pod to get rid of this issue. Could you please explain what exactly is happening and how to fix it ? Thanks. 

Regards,

Arun

On Tue, Apr 11, 2023 at 10:42 AM Aruna Manivannan <aruna.manivannan@gmail.com> wrote:

Hi Yogesh, 

Thanks for your response. Will use it and let you know the result. Thank you again.

Regards,

Arun

On Tue, Apr 11, 2023 at 1:58 AM Yogesh Mahajan <yogesh.mahajan@enterprisedb.com> wrote:

Hi Aruna,

You can use 'misc/ping' to monitor health checks.

E.g.

pgadmin:    image: dpage/pgadmin4    healthcheck:      test: ["CMD", "wget", "-O", "-", "http://localhost:80/misc/ping"]

Thanks,

Yogesh Mahajan

EnterpriseDB

On Tue, Apr 11, 2023 at 2:20 AM Aruna Manivannan <aruna.manivannan@gmail.com> wrote:

Hi Team, 

This is Arun. Could you please help me with setting up a Liveness & Readiness probe for pgadmin4 6.17 which is running in the Kubernetes cluster as Statefulset. I mean, is there any URL endpoint which I can configure to hit? Thanks. 

Regards,

Arun

--

Dave Page
Blog: https://pgsnake.blogspot.com
Twitter: @pgsnake

EDB: https://www.enterprisedb.com