Getting "Elastic did not load properly. Check the server output for more information." when accessing through gatekeeper

Question

I'm using a Gatekeeper (quay.io/keycloak/keycloak-gatekeeper:10.0.0) for SSO with keycloak (docker.io/jboss/keycloak:11.0.2) on Kibana (docker.elastic.co/kibana/kibana-oss:7.10.1)

My gatekeeper config:

discovery-url: http://{{ index .Values.gatekeeper.config "keycloak-url" }}/auth/realms/{{ .Values.gatekeeper.config.realm }}
skip-openid-provider-tls-verify: true
client-id: {{ index .Values.gatekeeper.config "client-id" }}
client-secret: {{ index .Values.gatekeeper.config "client-secret" }}
listen: :3000
enable-refresh-tokens: true
enable-default-deny: true
enable-session-cookies: true
tls-cert:
tls-private-key:
redirection-url: http://{{ index .Values.gatekeeper.config "redirection-url" }}:{{ .Values.service.nodePort }}
secure-cookie: false
encryption-key: {{ index .Values.gatekeeper.config "encryption-key" }}
upstream-url: http://127.0.0.1:5601/
forbidden-page: /html/access-forbidden.html
upstream-keepalives: true
resources:
- uri: /*

I'm able to get to the keycloak login page but after successful login (and after Kibana loader has finished loading) i get an error page:

from kibana logs i see:

{"type":"log","@timestamp":"2020-12-19T15:50:58Z","tags":["debug","connection","econnreset"],"pid":8,"message":"ECONNRESET: Socket was closed by the client (probably the browser) before

it could be read completely"} {"type":"log","@timestamp":"2020-12-19T15:51:00Z","tags":["debug","connection","econnreset"],"pid":8,"message":"ECONNRESET: Socket was closed by the client (probably the browser) before it could be read completely"} {"type":"log","@timestamp":"2020-12-19T15:51:01Z","tags":["debug","connection","econnreset"],"pid":8,"message":"ECONNRESET: Socket was closed by the client (probably the browser) before it could be read completely"}

Did anyone eve encountered this before? How can i fix this?

Thanks

Answer 1

For me, I disabled the adBlocker or similar and that's right.

Answer 2

In our case, Kibana is behind a Nginx LB and we've had that issue because the on DDOS protection, burst size was too small. I've increased that value from 10 to 120.

limit_req zone=limit_req_kibana.website burst=120 nodelay;