forked from grafana/agent
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
clean up the docker compose example environment (grafana#1456)
The various docker-compose.*.yaml files have been replaced with a single docker-compose.yaml file. By default, this file will spawn and expose ports for Loki, Grafana, Tempo, and Cortex so telemetry data can be stored and visualized. This allows developers to bring up dependencies needed for testing a local Grafana Agent process, useful for validating code changes. Version 3 of Docker Compose is used for profile support, allowing docker-compose.yaml to optionally run extra services: * The `agent` profile will also run a Grafana Agent that collects metrics from itself and sends them to Cortex. * The `integrations` profile will run example services that integrations can target, such as MySQL and Consul. Each of these `integrations` services can also be enabled independently by using the profile name matching the relevant integration name (i.e., the `mysqld_exporter` profile enables the MySQL service). All integration services will expose their ports for a local agent to connect to.
- Loading branch information
Showing
17 changed files
with
351 additions
and
825 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,74 +1,56 @@ | ||
# Example | ||
|
||
This directory contains an example deployment of the Grafana Agent with | ||
the following components: | ||
This directory contains a Docker Compose 3 environment that can be used to test | ||
Grafana Agent. | ||
|
||
1. Cortex to store metrics | ||
2. Grafana Agent to collect metrics | ||
3. Grafana to visualize metrics | ||
4. Avalanche to load test the Agent. | ||
By default, the following services are exposed: | ||
|
||
This example is used for seeing how a single instance of the Agent performs | ||
under moderate load; the Docker Compose configuration as present in this | ||
directory will generate roughly 90,000 metrics. | ||
1. Cortex for storing metrics (localhost:9009) | ||
2. Grafana for visualizing telemetry (localhost:3000) | ||
3. Loki for storing logs (localhost:3100) | ||
4. Tempo for storing traces (localhost:3200) | ||
5. Avalanche for a sample /metrics endpoint to scrape (localhost:9001). | ||
|
||
To get started, run the following from this directory: | ||
Run the following to bring up the environment: | ||
|
||
``` | ||
docker-compose up -d | ||
``` | ||
|
||
This will create Cortex storage in `/tmp/cortex` and the agent WAL in | ||
`/tmp/agent`; you may want to delete those directories after you are done | ||
testing. | ||
By default, the Docker Compose environment doesn't include a Grafana Agent | ||
container. This lets you test the agent externally, especially useful when | ||
validating code changes. You can enable the included Grafana Agent by passing | ||
`agent` to the profiles list: `docker compose --profile=agent up -d`. When | ||
running, the Agent exposes its HTTP endpoint at localhost:12345. | ||
|
||
Once the containers are running, a Grafana instance will be exposed at | ||
`http://localhost:3000` with Cortex as the only datasource. You should shortly | ||
see metrics in Cortex that are sent from the agent. Agent operational dashboards | ||
are included in the deployment. | ||
The Docker Compose environment heavily relies on profiles to enable optional | ||
features. You can pass multiple profiles by separating them by comma: | ||
`docker compose --profile=agent,integrations up -d`. | ||
|
||
The Agent is exposed on the host at `http://localhost:12345`. | ||
## Running Integrations | ||
|
||
## Hacking on the Example | ||
The Docker Compose environment includes example services to point integrations | ||
at (e.g., mysql, redis, consul, etc.). You can run all integrations using the | ||
`integrations` profile, or run individual services with a profile name matching | ||
the integration (i.e., enabling the `dnsmasq_exporter` profile to specifically | ||
enable the dnsmasq service). | ||
|
||
The reduced memory requirements is a critical feature of the Agent, and | ||
the example provides a good launching point to end-to-end test and validate | ||
the usage. | ||
Enabling specific integration profiles is useful when you only want to test a | ||
single integration, as the `integrations` profile can be resource intensive. | ||
|
||
To build the image locally, run `make agent-image` at the root of this | ||
repository. | ||
## Visualizing | ||
|
||
To get a memory profile, you can use `pprof` against the Agent: | ||
Grafana is exposed at `http://localhost:3000`, and includes some useful | ||
dashboards: | ||
|
||
``` | ||
go tool pprof -http=:6060 http://localhost:12345/debug/pprof/heap?debug=1 | ||
``` | ||
|
||
Useful one-off queries to run once everything is up: | ||
|
||
1. `agent_wal_storage_active_series`: How many series are active in the WAL | ||
2. `cortex_ingester_memory_series`: How many series are active in Cortex. | ||
Should be equal to the previous metric. | ||
3. `go_memstats_heap_inuse_bytes{container="agent"} / 1e6`: Current memory | ||
usage of agent in megabytes. | ||
4. `max by (container,instance,job) | ||
(avg_over_time(go_memstats_heap_inuse_bytes[10m])) / 1e6`: Current memory | ||
usage of the agent and Cortex averaged out from the last 10 minutes. | ||
* The `Agent` dashboard gives a very high-level overview of running agents. | ||
|
||
## Scraping Service | ||
|
||
A [Scraping Service](https://grafana.com/docs/agent/latest/scraping-service/) cluster example is also | ||
present in this directory. To start it, run: | ||
|
||
``` | ||
docker-compose -f docker-compose.scraping-service.yaml up -d | ||
``` | ||
* The `Agent Prometheus Remote Write` dashboard visualizes the current state of | ||
writing metrics to Cortex. | ||
|
||
The Scraping Service cluster will start with an empty KV store for instance | ||
configurations. A set of instance configuration files is provided in | ||
[`agent/instance-configs`](./agent/instance-configs) that will work with the | ||
Docker Compose example. | ||
* The `Agent Tracing Pipeline` dashbaord visualizes the current state of the | ||
tracing pipeline (if spans are being processed). | ||
|
||
The `agentctl` tool provided by this repository can automatically deploy | ||
those configs for you; run `agentctl config-sync -a http://localhost:12345 ./agent/instance-configs` | ||
from this directory to do so. | ||
* The `Agent Operational` dashboard shows resource consumption of Grafana | ||
Agent. Not all panels will have data here, as they rely on metrics from other | ||
sources (i.e., cAdvisor). |
23 changes: 0 additions & 23 deletions
23
example/docker-compose/agent/config/agent-integrations-next.yaml
This file was deleted.
Oops, something went wrong.
This file was deleted.
Oops, something went wrong.
62 changes: 0 additions & 62 deletions
62
example/docker-compose/agent/config/agent-scraping-service.yaml
This file was deleted.
Oops, something went wrong.
Oops, something went wrong.