From 0be3bf8eda98bb40fa881e3d1002bb686977d20f Mon Sep 17 00:00:00 2001
From: Andras Ferencz-Szabo <andrasferenczszabo@monzo.com>
Date: Thu, 17 Oct 2019 13:12:35 +0100
Subject: [PATCH] Add a new prometheus.ready_timeout CLI option to the sidecar

The current timeout while the Prometheus instance is starting up is a constant 10 minutes.
As reading the WAL can take a long time we would like to set a custom timeout value, so the Thanos
sidecar container is not erroring out every 10 minutes.
---
 CHANGELOG.md               |  4 ++++
 cmd/thanos/sidecar.go      | 11 +++++++----
 docs/components/sidecar.md |  3 +++
 3 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index dc2db529d9c..c7c45a703b1 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -11,6 +11,10 @@ We use *breaking* word for marking changes that are not backward compatible (rel
 
 ## Unreleased
 
+### Added
+
+- [#1660](https://github.com/thanos-io/thanos/pull/1660) Add a new `--prometheus.ready_timeout` CLI option to the sidecar to set how long to wait until Prometheus starts up.
+
 ## [v0.8.1](https://github.com/thanos-io/thanos/releases/tag/v0.8.1) - 2019.10.14
 
 ### Fixed
diff --git a/cmd/thanos/sidecar.go b/cmd/thanos/sidecar.go
index 403bce843e0..dfc2b4ac56b 100644
--- a/cmd/thanos/sidecar.go
+++ b/cmd/thanos/sidecar.go
@@ -32,8 +32,6 @@ import (
 	"gopkg.in/alecthomas/kingpin.v2"
 )
 
-const waitForExternalLabelsTimeout = 10 * time.Minute
-
 func registerSidecar(m map[string]setupFunc, app *kingpin.Application) {
 	cmd := app.Command(component.Sidecar.String(), "sidecar for Prometheus server")
 
@@ -43,6 +41,9 @@ func registerSidecar(m map[string]setupFunc, app *kingpin.Application) {
 	promURL := cmd.Flag("prometheus.url", "URL at which to reach Prometheus's API. For better performance use local network.").
 		Default("http://localhost:9090").URL()
 
+	promReadyTimeout := cmd.Flag("prometheus.ready_timeout", "Maximum time to wait for the Prometheus instance to start up").
+		Default("10m").Duration()
+
 	dataDir := cmd.Flag("tsdb.path", "Data directory of TSDB.").
 		Default("./data").String()
 
@@ -81,6 +82,7 @@ func registerSidecar(m map[string]setupFunc, app *kingpin.Application) {
 			*clientCA,
 			*httpBindAddr,
 			*promURL,
+			*promReadyTimeout,
 			*dataDir,
 			objStoreConfig,
 			rl,
@@ -102,6 +104,7 @@ func runSidecar(
 	clientCA string,
 	httpBindAddr string,
 	promURL *url.URL,
+	promReadyTimeout time.Duration,
 	dataDir string,
 	objStoreConfig *extflag.PathOrContent,
 	reloader *reloader.Reloader,
@@ -268,7 +271,7 @@ func runSidecar(
 		g.Add(func() error {
 			defer runutil.CloseWithLogOnErr(logger, bkt, "bucket client")
 
-			extLabelsCtx, cancel := context.WithTimeout(ctx, waitForExternalLabelsTimeout)
+			extLabelsCtx, cancel := context.WithTimeout(ctx, promReadyTimeout)
 			defer cancel()
 
 			if err := runutil.Retry(2*time.Second, extLabelsCtx.Done(), func() error {
@@ -277,7 +280,7 @@ func runSidecar(
 				}
 				return nil
 			}); err != nil {
-				return errors.Wrapf(err, "aborting as no external labels found after waiting %s", waitForExternalLabelsTimeout)
+				return errors.Wrapf(err, "aborting as no external labels found after waiting %s", promReadyTimeout)
 			}
 
 			var s *shipper.Shipper
diff --git a/docs/components/sidecar.md b/docs/components/sidecar.md
index 60699f125fb..a1d1a92112c 100644
--- a/docs/components/sidecar.md
+++ b/docs/components/sidecar.md
@@ -116,6 +116,9 @@ Flags:
       --prometheus.url=http://localhost:9090
                                  URL at which to reach Prometheus's API. For
                                  better performance use local network.
+      --prometheus.ready_timeout=10m
+                                 Maximum time to wait for the Prometheus
+                                 instance to start up
       --tsdb.path="./data"       Data directory of TSDB.
       --reloader.config-file=""  Config file watched by the reloader.
       --reloader.config-envsubst-file=""