SCALRCORE-35910 Helm chart config for agent-job #100
Conversation
mermoldy
changed the title
mermoldy
changed the title
mermoldy
force-pushed
the
SCALRCORE-35910
branch
from
9852e34 to
79afcde
Compare
mermoldy
force-pushed
the
SCALRCORE-35910
branch
from
233402f to
f4c26ea
Compare
mermoldy
force-pushed
the
SCALRCORE-35910
branch
from
599b483 to
88ee96c
Compare
mermoldy
changed the title
| - name: SCALR_URL | ||
| value: {{ .Values.agent.url | quote }} |
There was a problem hiding this comment.
@mermoldy Навіщо нам SCALR_URL? зараз домен є вже у токені
There was a problem hiding this comment.
@penja that’s optional configuration - in case the account migrates to a different URL, or the user wants to route requests through some proxy (there was simply no reason to remove it and introdude BC for old tokens)
| - name: SCALR_AGENT_DATA_DIR | ||
| value: {{ .Values.agent.dataDir | quote }} | ||
| - name: SCALR_AGENT_CACHE_DIR | ||
| value: {{ .Values.agent.cacheDir | quote }} |
There was a problem hiding this comment.
@mermoldy У нас є кейси копіювати кеш у датадиректорію. Як це буде працювати? Можливо нам не варто розділяти ці речі окремо?
There was a problem hiding this comment.
@penja the main point is to separate workspace (per-run) and cache (global) volumes - https://github.com/Scalr/agent-helm/tree/SCALRCORE-35910/charts/agent-job#volumes. Thats give us ability to use NFS only for the cache, and keep faster ephemeral volume for the run data directories The cache directory path by default (when not explictly specified) is equal to the data directory path to avoid unnecessary data migrations.
No description provided.