Skip to content

This charmed operator automates the operational procedures of running Loki, an open-source logging stack, on Kubernetes

License

Notifications You must be signed in to change notification settings

bencekov/loki-k8s-operator

 
 

Repository files navigation

Loki Charmed Operator for K8s

CharmHub Badge Release to Edge Release Libraries Discourse Status

Description

Loki is an open-source fully-featured logging system. This Loki charmed operator handles installation, scaling, configuration, optimisation, networking, service mesh, observability, and Day 2 operations specific to Loki using Juju and the Charmed Operator Lifecycle Manager (OLM).

On the principle that an operator should "do one thing and do it well", this operator drives Loki application only. However, it can be composed with other operators to deliver a complex application or service. Because operators package expert knowledge in a reusable and shareable form, they hugely simplify software management and operations.

Getting started

Basic deployment

Create a Juju model for your operator, say "observability"

juju add-model observability

The Loki Charmed Operator may be deployed using the Juju command line in a quite very way.

juju deploy loki-k8s --channel=stable

Checking deployment status

Once the Charmed Operator is deployed, the status can be checked running:

juju status --color --relations
$ juju status --color --relations

Model          Controller           Cloud/Region        Version  SLA          Timestamp
observability  charm-dev-batteries  microk8s/localhost  3.0.2    unsupported  10:39:52-03:00

App       Version  Status   Scale  Charm     Channel  Rev  Address         Exposed  Message
loki-k8s  2.4.1    waiting      1  loki-k8s  stable    47  10.152.183.195  no       waiting for container

Unit         Workload  Agent  Address      Ports  Message
loki-k8s/0*  active    idle   10.1.36.115

Loki HTTP API

Loki Charmed Operator exposes its HTTP API over port 3100.

Example 1 - Get Loki version:

/loki/api/v1/status/buildinfo exposes the build information in a JSON object. The fields are version, revision, branch, buildDate, buildUser, and goVersion.

loki_ip=$(juju status loki-k8s/0 | grep "loki-k8s/0" | awk '{print $4}')

curl http://$loki_ip:3100/loki/api/v1/status/buildinfo
{"version":"2.4.1","revision":"f61a4d261","branch":"HEAD","buildUser":"root@39a6e600b2df","buildDate":"2021-11-08T13:09:51Z","goVersion":""}

Example 2 - Send logs entries to Loki with curl:

/loki/api/v1/push is the endpoint used to send log entries to Loki. The default behavior is for the POST body to be a snappy-compressed protobuf message. Alternatively, if the Content-Type header is set to application/json, a JSON post body can be sent in the following format:

loki_ip=$(juju status loki-k8s/0 | grep "loki-k8s/0" | awk '{print $4}')

curl -v -H "Content-Type: application/json" -XPOST -s "http://$loki_ip:3100/loki/api/v1/push" --data-raw \
  '{"streams": [{ "stream": { "foo": "bar2" }, "values": [ [ "1570818238000000000", "fizzbuzz" ] ] }]}'

Example 3 - Send logs entries to Loki with Promtail:

Promtail is an agent which ships the contents of local logs to Loki. It is usually deployed to every machine that has applications needed to be monitored.

It primarily:

  • Discovers targets
  • Attaches labels to log streams
  • Pushes them to the Loki instance.

Currently, Promtail can tail logs from two sources: local log files and the systemd journal (on AMD64 machines only).

To set up a Promtail instance to work with Loki Charmed Operator please refer to Configuring Promtail documentation. Anyway the most important part is the clients section in Promtail config file, for instance:

clients:
  - url: http://<LOKI_ADDRESS>:3100/loki/api/v1/push

Relations

Overview

Relations provide a means to integrate applications and enable a simple communications channel. Loki Charmed Operator supports the following:

Provides

Logging

  logging:
    interface: loki_push_api

Loki Charmed Operator may receive logs from any charm that supports the loki_push_api relation interface.

Let's say that we have a Charmed Operator that implements the other side (requires) of the relation, for instance Zinc After deploying this charm, we can relate Loki and Zinc through loki_push_api relation interface:

juju relate zinc-k8s loki-k8s

And verify the relation between both charms is created:

$ juju status --relations

Model          Controller           Cloud/Region        Version  SLA          Timestamp
observability  charm-dev-batteries  microk8s/localhost  3.0.2    unsupported  10:57:01-03:00

App       Version  Status  Scale  Charm     Channel  Rev  Address         Exposed  Message
loki-k8s  2.4.1    active      1  loki-k8s  stable    47  10.152.183.168  no
zinc-k8s  0.3.5    active      1  zinc-k8s  stable    45  10.152.183.144  no

Unit         Workload  Agent      Address      Ports  Message
loki-k8s/0*  active    executing  10.1.36.79
zinc-k8s/0*  active    executing  10.1.36.123

Relation provider  Requirer          Interface      Type     Message
loki-k8s:logging   zinc-k8s:logging  loki_push_api  regular

Once the relation is established, Zinc charm can start sending logs to Loki charm.

Grafana-source

  grafana-source:
    interface: grafana_datasource
    optional: true

The Grafana Charmed Operator aggregates logs obtained by Loki and provides a versatile dashboard to view these logs in configurable ways. Loki relates to Grafana over the grafana_datasource interface.

For example, let's say that we have already deployed the Grafana Charmed Operator in our observability model. The way to relate Loki and Grafana is again very simple:

juju relate grafana-k8s:grafana-source loki-k8s:grafana-source

And verify the relation between both charms is created:

$ juju status --relations

Model          Controller           Cloud/Region        Version  SLA          Timestamp
observability  charm-dev-batteries  microk8s/localhost  3.0.2    unsupported  11:09:08-03:00

App          Version  Status  Scale  Charm        Channel  Rev  Address         Exposed  Message
grafana-k8s  9.2.1    active      1  grafana-k8s  stable    52  10.152.183.40   no
loki-k8s     2.4.1    active      1  loki-k8s     stable    47  10.152.183.168  no

Unit            Workload  Agent  Address     Ports  Message
grafana-k8s/0*  active    idle   10.1.36.93
loki-k8s/0*     active    idle   10.1.36.79

Relation provider        Requirer                    Interface           Type     Message
grafana-k8s:grafana      grafana-k8s:grafana         grafana_peers       peer
loki-k8s:grafana-source  grafana-k8s:grafana-source  grafana_datasource  regular

Metrics-endopoint

  metrics-endpoint:
    interface: prometheus_scrape

This Loki Charmed Operator provides a metrics endpoint so a charm that implements the other side of the relation (requires), for instance Prometheus Charmed Operator can scrape Loki metrics.

For instance, let's say that we have already deployed the Prometheus Charmed Operator in our observability model. The way to relate Loki and Promethes is again very simple:

juju relate prometheus-k8s loki-k8s
$ juju status --relations

Model          Controller           Cloud/Region        Version  SLA          Timestamp
observability  charm-dev-batteries  microk8s/localhost  3.0.2    unsupported  11:25:19-03:00

App             Version  Status  Scale  Charm           Channel  Rev  Address         Exposed  Message
grafana-k8s     9.2.1    active      1  grafana-k8s     stable    52  10.152.183.40   no
loki-k8s        2.4.1    active      1  loki-k8s        stable    47  10.152.183.168  no
prometheus-k8s  2.33.5   active      1  prometheus-k8s  stable    79  10.152.183.144  no

Unit               Workload  Agent  Address     Ports  Message
grafana-k8s/0*     active    idle   10.1.36.93
loki-k8s/0*        active    idle   10.1.36.79
prometheus-k8s/0*  active    idle   10.1.36.84

Relation provider                Requirer                         Interface           Type     Message
grafana-k8s:grafana              grafana-k8s:grafana              grafana_peers       peer
loki-k8s:grafana-source          grafana-k8s:grafana-source       grafana_datasource  regular
loki-k8s:metrics-endpoint        prometheus-k8s:metrics-endpoint  prometheus_scrape   regular
prometheus-k8s:prometheus-peers  prometheus-k8s:prometheus-peers  prometheus_peers    peer

Grafana-dashboard

  grafana-dashboard:
    interface: grafana_dashboard

Loki Charmed Operator may send its own Dashboards to Grafana by using this relation.

After relating both charms this way:

juju relate grafana-k8s:grafana-dashboard loki-k8s:grafana-dashboard

You will be able to check that the relation is established, and see the Loki Dashboard in Grafana UI.

$ juju status --relations

Model          Controller           Cloud/Region        Version  SLA          Timestamp
observability  charm-dev-batteries  microk8s/localhost  3.0.2    unsupported  13:21:21-03:00

App             Version  Status  Scale  Charm           Channel  Rev  Address         Exposed  Message
grafana-k8s     9.2.1    active      1  grafana-k8s     stable    52  10.152.183.40   no
loki-k8s        2.4.1    active      1  loki-k8s        stable    47  10.152.183.168  no
prometheus-k8s  2.33.5   active      1  prometheus-k8s  stable    79  10.152.183.144  no

Unit               Workload  Agent  Address     Ports  Message
grafana-k8s/0*     active    idle   10.1.36.85
loki-k8s/0*        active    idle   10.1.36.97
prometheus-k8s/0*  active    idle   10.1.36.67

Relation provider                Requirer                         Interface           Type     Message
grafana-k8s:grafana              grafana-k8s:grafana              grafana_peers       peer
loki-k8s:grafana-dashboard       grafana-k8s:grafana-dashboard    grafana_dashboard   regular
loki-k8s:grafana-source          grafana-k8s:grafana-source       grafana_datasource  regular
loki-k8s:metrics-endpoint        prometheus-k8s:metrics-endpoint  prometheus_scrape   regular
prometheus-k8s:prometheus-peers  prometheus-k8s:prometheus-peers  prometheus_peers    peer

Requires

Alertmanager

Alertmanager receives alerts from Loki, aggregates and deduplicates them, then forwards them to specified targets. Loki Charmed Operator relates to Alertmanager over the alertmanager_dispatch interface.

Let's assume the we have already deployed Alertmanager Charmed Operator in our obsevability model, and relate it with Loki:

juju relate alertmanager-k8s loki-k8s

We can check the relation is established:

juju status --color --relations
$ juju status --relations
Model          Controller           Cloud/Region        Version  SLA          Timestamp
observability  charm-dev-batteries  microk8s/localhost  3.0.2    unsupported  13:27:18-03:00

App               Version  Status   Scale  Charm             Channel  Rev  Address         Exposed  Message
alertmanager-k8s  0.23.0   waiting      1  alertmanager-k8s  stable    36  10.152.183.56   no       waiting for container
grafana-k8s       9.2.1    active       1  grafana-k8s       stable    52  10.152.183.40   no
loki-k8s          2.4.1    active       1  loki-k8s          stable    47  10.152.183.168  no
prometheus-k8s    2.33.5   active       1  prometheus-k8s    stable    79  10.152.183.144  no

Unit                 Workload  Agent      Address      Ports  Message
alertmanager-k8s/0*  active    idle       10.1.36.113
grafana-k8s/0*       active    idle       10.1.36.85
loki-k8s/0*          active    executing  10.1.36.97
prometheus-k8s/0*    active    idle       10.1.36.67

Relation provider                Requirer                         Interface              Type     Message
alertmanager-k8s:alerting        loki-k8s:alertmanager            alertmanager_dispatch  regular
alertmanager-k8s:replicas        alertmanager-k8s:replicas        alertmanager_replica   peer
grafana-k8s:grafana              grafana-k8s:grafana              grafana_peers          peer
loki-k8s:grafana-dashboard       grafana-k8s:grafana-dashboard    grafana_dashboard      regular
loki-k8s:grafana-source          grafana-k8s:grafana-source       grafana_datasource     regular
loki-k8s:metrics-endpoint        prometheus-k8s:metrics-endpoint  prometheus_scrape      regular
prometheus-k8s:prometheus-peers  prometheus-k8s:prometheus-peers  prometheus_peers       peer

Ingress

  ingress:
    interface: ingress_per_unit
    limit: 1

Interactions with the Loki charm can not be assumed to originate within the same Juju model, let alone the same Kubernetes cluster, or even the same Juju cloud. Hence the Loki charm also supports an Ingress relation. There are multiple use cases that require an ingress, in particular

  • Querying the Loki HTTP API endpoint across network boundaries.
  • Self monitoring of Loki that must happen across network boundaries to ensure robustness of self monitoring.
  • Supporting the Loki push API.

Loki typical needs a "per unit" Ingress. This per unit ingress is necessary since Loki exposes it loki push api endpoint on a per unit basis. A per unit ingress relation is available in the traefik-k8s charm and this Loki charm does support that relation over ingress_per_unit interface.

Let's assume the we have already deployed Traefik Charmed Operator in our obsevability model, and relate it with Loki:

juju relate traefik-k8s loki-k8s

We can check the relation is established:

juju status --color --relations
$ juju status --relations
Model          Controller           Cloud/Region        Version  SLA          Timestamp
observability  charm-dev-batteries  microk8s/localhost  3.0.2    unsupported  15:46:43-03:00

App               Version  Status  Scale  Charm             Channel  Rev  Address         Exposed  Message
alertmanager-k8s  0.23.0   active      1  alertmanager-k8s  stable    36  10.152.183.56   no
grafana-k8s       9.2.1    active      1  grafana-k8s       stable    52  10.152.183.40   no
loki-k8s          2.4.1    active      1  loki-k8s          stable    47  10.152.183.168  no
prometheus-k8s    2.33.5   active      1  prometheus-k8s    stable    79  10.152.183.144  no
traefik-k8s                active      1  traefik-k8s       stable    93  192.168.122.10  no

Unit                 Workload  Agent      Address      Ports  Message
alertmanager-k8s/0*  active    idle       10.1.36.95
grafana-k8s/0*       active    executing  10.1.36.121
loki-k8s/0*          active    idle       10.1.36.116
prometheus-k8s/0*    active    idle       10.1.36.80
traefik-k8s/0*       active    idle       10.1.36.122

Relation provider                Requirer                         Interface              Type     Message
alertmanager-k8s:alerting        loki-k8s:alertmanager            alertmanager_dispatch  regular
alertmanager-k8s:replicas        alertmanager-k8s:replicas        alertmanager_replica   peer
grafana-k8s:grafana              grafana-k8s:grafana              grafana_peers          peer
loki-k8s:grafana-dashboard       grafana-k8s:grafana-dashboard    grafana_dashboard      regular
loki-k8s:grafana-source          grafana-k8s:grafana-source       grafana_datasource     regular
loki-k8s:metrics-endpoint        prometheus-k8s:metrics-endpoint  prometheus_scrape      regular
prometheus-k8s:prometheus-peers  prometheus-k8s:prometheus-peers  prometheus_peers       peer
traefik-k8s:ingress-per-unit     loki-k8s:ingress                 ingress_per_unit       regular

OCI Images

Every release of the Loki Charmed Operator uses the latest stable version of grafana/loki at the time of release.

Official Documentation

For further details about Loki configuration and usage, please refer to Grafana Loki Documentation

About

This charmed operator automates the operational procedures of running Loki, an open-source logging stack, on Kubernetes

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%