jamesread/prometheus-gmail-exporter

Question: How do I skip the expiry on a headless server?

Opened this issue · 6 comments

Kiara0107 commented

As advised in the READ.ME I have created the login_cookie.dat on my local Windows host (with internet). I moved that file to the headless linux server where I want to run the docker container. It's working for 1 hour, then the file is expired. How do I prevent expiry?

login_cookie.dat

{
  "token": "[[very-long-token]]"
, "refresh_token": "[[long-token]]"
, "token_uri": "https://oauth2.googleapis.com/token"
, "client_id": "[[ID]].apps.googleusercontent.com"
, "client_secret": "[[secret]]"
, "scopes": "https://www.googleapis.com/auth/gmail.readonly "
, "expiry": "2023-08-24T10:59:26.026612Z"
}

client_secret.json

{"web":
 {
  "client_id":"[[ID]].apps.googleusercontent.com"
  ,"project_id":"mailbox"
  ,"auth_uri":"https://accounts.google.com/o/oauth2/auth"
  ,"token_uri":"https://oauth2.googleapis.com/token"
  ,"auth_provider_x509_cert_url":"https://www.googleapis.com/oauth2/v1/certs"
  ,"client_secret":"[[secret]]"
 }
}

Linux Ubuntu 20.04.6 LTS (GNU/Linux 5.4.0-156-generic x86_64)

docker run -d -p 9117:8080 \
      -v /etc/gmail/:/root/.prometheus-gmail-exporter/ \
      --restart unless-stopped \
      --network prometheusnet \
      --name gmail-exporter gmail-exporter:latest \
      --updateDelaySeconds 60 \
      --daemonize \
      --labels INBOX

Logfile after 1 hour shows:

INFO:root:Updating gmail metrics - started
INFO:google_auth_httplib2:Refreshing credentials due to a 401 response. Attempt 1/2.
ERROR:root:Error: ('invalid_scope: Some requested scopes were invalid. {invalid=[a, c, d, e, g, h, i, l, m, ., n, /, o, p, r, s, t, u, w, y, :]}', {'error': 'invalid_scope', 'error_description': 'Some requested scopes were invalid. {invalid=[a, c, d, e, g, h, i, l, m, ., n, /, o, p, r, s, t, u, w, y, :]}', 'error_uri': 'https://developers.google.com/identity/protocols/oauth2'})
INFO:root:Updating gmail metrics - complete

I'm probably missing a simple step, can you help me out? :)

lapo-luchini commented

On a 401 would also be nice to delete metrics and stop exporting (outdated) values to Prometheus, so that we can notice this is happening from the graph.

lapo-luchini commented

For me it works barely for 20 minutes after authentication. :(

[2024-05-14 12:21:58] INFO:googleapiclient.discovery_cache:file_cache is only supported with oauth2client<4.0.0
[2024-05-14 12:21:58] INFO:root:Got gmail client successfully
[2024-05-14 12:21:58] INFO:root:Prometheus started on port 9137
[2024-05-14 12:21:58] INFO:root:Updating gmail metrics - started
[2024-05-14 12:21:58] INFO:root:Getting metadata about labels
[2024-05-14 12:21:58] INFO:root:Using labels: ['INBOX']
[2024-05-14 12:21:58] INFO:root:Updating gmail metrics - complete
[2024-05-14 12:31:58] INFO:root:Updating gmail metrics - started
[2024-05-14 12:31:59] INFO:root:Updating gmail metrics - complete
[2024-05-14 12:41:59] INFO:root:Updating gmail metrics - started
[2024-05-14 12:42:00] INFO:root:Updating gmail metrics - complete
[2024-05-14 12:52:00] INFO:root:Updating gmail metrics - started
[2024-05-14 12:52:00] ERROR:root:Error: ('invalid_scope: Some requested scopes were invalid. …', 'error_uri': 'https://developers.google.com/identity/protocols/oauth2'})
jamesread commented

Hey @lapo-luchini , gmail / google is kinda shitty with their APIs - things seem to randomly change and aren't versioned great. It does look like some scopes were renamed which I need to look into - see #32 (comment) .

I agree that this app should expose this as an error to prom - I don't know what is the idiomatic way to do this in Prometheus, so I'll start some research on that.

jamesread commented

@Kiara0107 - the app needs to constantly refresh login_cookie.dat - but it should re-login automatically without you having to open a browser all the time. Sorry that you have been waiting so long for me to reply.

@lapo-luchini - the error message you're posting isn't logging you out, but rather, you just have some bad scopes configured. I had some old documentation in the readme about scopes - which I've updated just now. Can you re-check your scopes? If you remove the invalid ones your error should disapear.

lapo-luchini commented

the error message you're posting isn't logging you out, but rather, you just have some bad scopes configured.

Oh, okay! I did need to re-login when stopping and immediately restarting the app, so I tought it was so.
I will try agani this evening. Thanks for the feedback!

jamesread commented

Make sure you update your container image - the version posted to docker hub was very old (I just deleted it), the version on ghcr.io is good 😊