[BUG]

[fbossiere]

If you have problems with the installation please use our community forum

Describe the bug
Hi,

I've just installed Leantime with this helmchart and I notice some post requests actions take incredibly long time (like create a project for example, or edit a task which lasts 30 second). I see no error log at all nor any indication.
I tested with both internal MariaDB or external one and the result is the same.

Here is a log example :
100.64.1.181 - - [19/Sep/2020:08:25:01 +0000] "GET /api/users?profileImage=currentUser HTTP/2.0" 302 0 "https://leantime.polynom.io/tickets/showTicket/12" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/85.0.4183.102 Safari/537.36"
100.64.1.181 - - [19/Sep/2020:08:25:01 +0000] "PATCH /api/sessions HTTP/2.0" 200 0 "https://leantime.polynom.io/tickets/showTicket/12" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/85.0.4183.102 Safari/537.36"
100.64.1.181 - - [19/Sep/2020:08:25:01 +0000] "PATCH /api/sessions HTTP/2.0" 200 0 "https://leantime.polynom.io/tickets/showTicket/12" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/85.0.4183.102 Safari/537.36"
100.64.1.181 - - [19/Sep/2020:08:25:37 +0000] "POST /tickets/showTicket/12 HTTP/2.0" 302 0 "https://leantime.polynom.io/tickets/showTicket/12" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/85.0.4183.102 Safari/537.36"

It looks like the GET queries get handled smoothly by the backend, but not the POST queries which looks redirected (code 302)

Can you help me understand what's going on ?

To Reproduce
Steps to reproduce the behavior:

1. Deploy Leantime with helm chart from https://github.com/gissilabs/charts/
2. Click create a project

Expected behavior
The time for the query to compute can be more than 30 seconds.

Leantime Version
2.1.4

Server
Apache (given in the docker)

[marcelfolaron]

@sgissi would you be able to help out?

[sgissi]

Hi, I have just deployed a new default install on my environment and I have no slowness at all. I also created new projects in my production deployment with Ceph as persistent storage without any issues. Can you give us more details about your cluster and how do you access the service (ingress controller, kubectl proxy, etc.)? If you deploy without any persistent storage (just "helm install leantime gissilabs/leantime") do you also have the same issue?

[fbossiere]

Sure, I have a K8S cluster hosted on Scaleway, I use a Kong loadbalancer to redirect my requests to a NodePort service which serves the Leantime Deployment. I have many more deployments on the cluster which works well.

[sgissi]

Thanks for that. Being exactly 30s makes it sounds like a timeout somewhere. @jjensen90 I could not find a way to increase log verbosity or trace internal calls. I cannot reproduce the issue on my cluster.

[sgissi]

@pacobuenaparte Can you try to access the Leantime pod directly using "kubectl port-forward"? Just to eliminate variables. If you can share the custom values you put into the chart (minus the sensitive parts), that would also help.

[fbossiere]

The same error occurs in port-forward mode

Note that the S3 bucket I provided do not seem to work. Could it be a call to the bucket which fails in a timeout ?

[fbossiere]

This is the chart

affinity: {}
externalDatabase:
  database: leantime
  enabled: true
  host: xxxxxxxxxxxxxx
  password: "xxxxxxxxxxxxxxxxxxx"
  user: "leantime_user"
fullnameOverride: ""
image:
  pullPolicy: IfNotPresent
  repository: leantime/leantime
  tag: ""
imagePullSecrets: []
ingress:
  annotations: {}
  enabled: false
  host: ""
  tls: []
ingressRoute:
  enabled: false
  entrypoints:
  - websecure
  host: ""
  tls: []
internalDatabase:
  enabled: false
leantime:
  color: ""
  language: en-US
  logo: ""
  name: leantime-xxxxxxx
  s3:
    bucket: leantime-xxxxxxx-bucket
    enabled: true
    key: xxxxxxxxxxxxxxxxxxxxxxxxx
    region: eu-central-1
    secret: xxxxxxxxxxxxxxxxxxx
  sessionSalt: ""
  smtp:
    autoTLS: true
    enabled: true
    from: leantime@xxxxxxxxxxxx.io
    host: email-smtp.eu-central-1.amazonaws.com
    password: xxxxxxxxxxxxxxxxxxxxxx
    port: 465
    secureProtocol: tls
    user: xxxxxxxxxxxxxxxx
  url: https://leantime.xxxxxxxxxxx.io
nameOverride: ""
nodeSelector: {}
persistence:
  accessMode: ReadWriteOnce
  enabled: true
  size: 1Gi
podAnnotations: {}
podSecurityContext: {}
replicaCount: 1
resources: {}
securityContext: {}
service:
  externalTrafficPolicy: Cluster
  type: NodePort
serviceAccount:
  annotations: {}
  create: true
  name: ""
strategy: {}
tolerations: []

[fbossiere]

Ok I solved it : my smtp credentials were wrong. I guess when I disabled the smtp in my chart, the slowness went away. Can you tell me what is the best way to test my smtp credentials ?

Thank you,
Paco

[sgissi]

Glad to hear you found it! You should be able to teat SMTP credentials with a mail client. @jjensen90 @marcelfolaron Shouldn’t SMTP errors show in the log? I’ll run some tests here to confirm.