Awesome

cortex-tenant

Prometheus remote write proxy which marks timeseries with a Cortex/Mimir tenant ID based on labels.

Architecture

Overview

Cortex/Mimir tenants (separate namespaces where metrics are stored to and queried from) are identified by X-Scope-OrgID HTTP header on both writes and queries.

~~Problem is that Prometheus can't be configured to send this header~~ Actually in some recent version (year 2021 onwards) this functionality was added, but the tenant is the same for all jobs. This makes it impossible to use a single Prometheus (or an HA pair) to write to multiple tenants.

This software solves the problem using the following logic:

Receive Prometheus remote write
Search each timeseries for a specific label name and extract a tenant ID from its value. If the label wasn't found then it can fall back to a configurable default ID. If none is configured then the write request will be rejected with HTTP code 400
Optionally removes this label from the timeseries
Groups timeseries by tenant
Issues a number of parallel per-tenant HTTP requests to Cortex/Mimir with the relevant tenant HTTP header (X-Scope-OrgID by default)

Usage

Get rpm or deb for amd64 from the Releases page. For building see below.

HTTP Endpoints

GET /alive returns 200 by default and 503 if the service is shutting down (if timeout_shutdown setting is > 0)
POST /push receives metrics from Prometheus - configure remote write to send here

Configuration

The service can be configured by a config file and/or environment variables. Config file may be specified by passing -config CLI argument.

If both are used then the env vars have precedence (i.e. they override values from config). See below for config file format and corresponding env vars.

# Where to listen for incoming write requests from Prometheus
# env: CT_LISTEN
listen: 0.0.0.0:8080

# Profiling API, remove to disable
# env: CT_LISTEN_PPROF
listen_pprof: 0.0.0.0:7008

# Where to send the modified requests (Cortex/Mimir)
# env: CT_TARGET
target: http://127.0.0.1:9091/receive

# Whether to enable querying for IPv6 records
# env: CT_ENABLE_IPV6
enable_ipv6: false

# This parameter sets the limit for the count of outgoing concurrent connections to Cortex / Mimir.
# By default it's 64 and if all of these connections are busy you will get errors when pushing from Prometheus.
# If your `target` is a DNS name that resolves to several IPs then this will be a per-IP limit.
# env: CT_MAX_CONNS_PER_HOST
max_conns_per_host: 0

# Authentication (optional)
auth:
  # Egress HTTP basic auth -> add `Authentication` header to outgoing requests
  egress:
    # env: CT_AUTH_EGRESS_USERNAME
    username: foo
    # env: CT_AUTH_EGRESS_PASSWORD
    password: bar

# Log level
# env: CT_LOG_LEVEL
log_level: warn

# HTTP request timeout
# env: CT_TIMEOUT
timeout: 10s

# Timeout to wait on shutdown to allow load balancers detect that we're going away.
# During this period after the shutdown command the /alive endpoint will reply with HTTP 503.
# Set to 0s to disable.
# env: CT_TIMEOUT_SHUTDOWN
timeout_shutdown: 10s

# Max number of parallel incoming HTTP requests to handle
# env: CT_CONCURRENCY
concurrency: 10

# Whether to forward metrics metadata from Prometheus to Cortex/Mimir
# Since metadata requests have no timeseries in them - we cannot divide them into tenants
# So the metadata requests will be sent to the default tenant only, if one is not defined - they will be dropped
# env: CT_METADATA
metadata: false

# If true response codes from metrics backend will be logged to stdout. This setting can be used to suppress errors
# which can be quite verbose like 400 code - out-of-order samples or 429 on hitting ingestion limits
# Also, those are already reported by other services like Cortex/Mimir distributors and ingesters
# env: CT_LOG_RESPONSE_ERRORS
log_response_errors: true

# Maximum duration to keep outgoing connections alive (to Cortex/Mimir)
# Useful for resetting L4 load-balancer state
# Use 0 to keep them indefinitely
# env: CT_MAX_CONN_DURATION
max_connection_duration: 0s

# Address where metrics are available
# env: CT_LISTEN_METRICS_ADDRESS
listen_metrics_address: 0.0.0.0:9090

# If true, then a label with the tenant’s name will be added to the metrics
# env: CT_METRICS_INCLUDE_TENANT
metrics_include_tenant: true

tenant:
  # Which label to look for the tenant information
  # env: CT_TENANT_LABEL
  label: tenant

  # List of labels examined for tenant information. If set takes precedent over `label`
  # env: CT_TENANT_LABEL_LIST
  label_list:
    - tenant
    - other_tenant

  # Whether to remove the tenant label from the request
  # env: CT_TENANT_LABEL_REMOVE
  label_remove: true
  
  # To which header to add the tenant ID
  # env: CT_TENANT_HEADER
  header: X-Scope-OrgID

  # Which tenant ID to use if the label is missing in any of the timeseries
  # If this is not set or empty then the write request with missing tenant label
  # will be rejected with HTTP code 400
  # env: CT_TENANT_DEFAULT
  default: foobar

  # Enable if you want all metrics from Prometheus to be accepted with a 204 HTTP code
  # regardless of the response from upstream. This can lose metrics if Cortex/Mimir is
  # throwing rejections.
  # env: CT_TENANT_ACCEPT_ALL
  accept_all: false

  # Optional prefix to be added to a tenant header before sending it to Cortex/Mimir.
  # Make sure to use only allowed characters:
  # https://grafana.com/docs/mimir/latest/configure/about-tenant-ids/
  # env: CT_TENANT_PREFIX
  prefix: foobar-

  # If true will use the tenant ID of the inbound request as the prefix of the new tenant id.
  # Will be automatically suffixed with a `-` character.
  # Example:
  #   Prometheus forwards metrics with `X-Scope-OrgID: Prom-A` set in the inbound request.
  #   This would result in the tenant prefix being set to `Prom-A-`.
  # https://grafana.com/docs/mimir/latest/configure/about-tenant-ids/
  # env: CT_TENANT_PREFIX_PREFER_SOURCE
  prefix_prefer_source: false

Prometheus configuration example

remote_write:
  - name: cortex_tenant
    url: http://127.0.0.1:8080/push

scrape_configs:
  - job_name: job1
    scrape_interval: 60s
    static_configs:
      - targets:
          - target1:9090
        labels:
          tenant: foobar

  - job_name: job2
    scrape_interval: 60s
    static_configs:
      - targets:
          - target2:9090
        labels:
          tenant: deadbeef

This would result in job1 metrics ending up in the foobar tenant in Cortex/Mimir and job2 in deadbeef.

Building

make build should create you an amd64 binary.

If you want deb or rpm packages then install FPM and then run make rpm or make deb to create the packages.

Containerization

To use the current container you need to overwrite the default configuration file, mount your configuration into to /data/cortex-tenant.yml.

You can overwrite the default config by starting the container with:

docker container run \
-v <CONFIG_LOCATION>:/data/cortex-tenant.yml \
ghcr.io/blind-oracle/cortex-tenant:1.6.1

... or build your own Docker image:

FROM ghcr.io/blind-oracle/cortex-tenant:1.6.1
ADD my-config.yml /data/cortex-tenant.yml

Deploy on Kubernetes

Using manifests

deploy/k8s/manifests directory contains the deployment, service and configmap manifest files for deploying this on Kubernetes. You can overwrite the default config by editing the configuration parameters in the configmap manifest.

kubectl apply -f deploy/k8s/manifests/cortex-tenant-deployment.yaml
kubectl apply -f deploy/k8s/manifests/cortex-tenant-service.yaml
kubectl apply -f deploy/k8s/manifests/config-file-configmap.yml

Using a Helm Chart

deploy/k8s/chart directory contains a chart for deploying this on Kubernetes.

helm repo add cortex-tenant https://blind-oracle.github.io/cortex-tenant
helm install cortex-tenant cortex-tenant/cortex-tenant

You can use deploy/k8s/chart/testing directory to test the deployment using helmfile.

helmfile -f deploy/k8s/chart/testing/helmfile.yaml template
helmfile -f deploy/k8s/chart/testing/helmfile.yaml apply

Updating chart version

helm package ./deploy/k8s/chart -d docs
TZ=UTC helm repo index ./docs