redis/example/otel/uptrace.yml

298 lines
8.4 KiB
YAML

##
## Uptrace configuration file.
## See https://uptrace.dev/get/config.html for details.
##
## You can use environment variables anywhere in this file, for example:
##
## foo: $FOO
## bar: ${BAR}
## baz: ${BAZ:default}
##
## To escape `$`, use `$$`, for example:
##
## foo: $$FOO_BAR
##
##
## ClickHouse database credentials.
##
ch:
# Connection string for ClickHouse database. For example:
# clickhouse://<user>:<password>@<host>:<port>/<database>?sslmode=disable
#
# See https://clickhouse.uptrace.dev/guide/golang-clickhouse.html#options
dsn: 'clickhouse://default:@clickhouse:9000/uptrace?sslmode=disable'
##
## A list of pre-configured projects. Each project is fully isolated.
##
projects:
# Conventionally, the first project is used to monitor Uptrace itself.
- id: 1
name: Uptrace
# Token grants write access to the project. Keep a secret.
token: project1_secret_token
pinned_attrs:
- service.name
- host.name
- deployment.environment
# Group spans by deployment.environment attribute.
group_by_env: false
# Group funcs spans by service.name attribute.
group_funcs_by_service: false
# Other projects can be used to monitor your applications.
# To monitor micro-services or multiple related services, use a single project.
- id: 2
name: My project
token: project2_secret_token
pinned_attrs:
- service.name
- host.name
- deployment.environment
# Group spans by deployment.environment attribute.
group_by_env: false
# Group funcs spans by service.name attribute.
group_funcs_by_service: false
##
## Create metrics from spans and events.
##
metrics_from_spans:
- name: uptrace.tracing.spans_duration
description: Spans duration (excluding events)
instrument: histogram
unit: microseconds
value: span.duration / 1000
attrs:
- span.system as system
- service.name as service
- host.name as host
- span.status_code as status
where: not span.is_event
- name: uptrace.tracing.spans
description: Spans count (excluding events)
instrument: counter
unit: 1
value: span.count
attrs:
- span.system as system
- service.name as service
- host.name as host
- span.status_code as status
where: not span.is_event
- name: uptrace.tracing.events
description: Events count (excluding spans)
instrument: counter
unit: 1
value: span.count
attrs:
- span.system as system
- service.name as service
- host.name as host
where: span.is_event
##
## To require authentication, uncomment the following section.
##
auth:
# users:
# - username: uptrace
# password: uptrace
# - username: admin
# password: admin
# # Cloudflare user provider: uses Cloudflare Zero Trust Access (Identity)
# # See https://developers.cloudflare.com/cloudflare-one/identity/ for more info.
# cloudflare:
# # The base URL of the Cloudflare Zero Trust team.
# - team_url: https://myteam.cloudflareaccess.com
# # The Application Audience (AUD) Tag for this application.
# # You can retrieve this from the Cloudflare Zero Trust 'Access' Dashboard.
# audience: bea6df23b944e4a0cd178609ba1bb64dc98dfe1f66ae7b918e563f6cf28b37e0
# # OpenID Connect (Single Sign-On)
# oidc:
# # The ID is used in API endpoints, for example, in redirect URL
# # `http://<uptrace-host>/api/v1/sso/<oidc-id>/callback`.
# - id: keycloak
# # Display name for the button in the login form.
# # Default to 'OpenID Connect'
# display_name: Keycloak
# # The base URL for the OIDC provider.
# issuer_url: http://localhost:8080/realms/uptrace
# # The OAuth 2.0 Client ID
# client_id: uptrace
# # The OAuth 2.0 Client Secret
# client_secret: ogbhd8Q0X0e5AZFGSG3m9oirPvnetqkA
# # Additional OAuth 2.0 scopes to request from the OIDC provider.
# # Defaults to 'profile'. 'openid' is requested by default and need not be specified.
# scopes:
# - profile
# # The OIDC UserInfo claim to use as the user's username.
# # Defaults to 'preferred_username'.
# claim: preferred_username
##
## Alerting rules for monitoring metrics.
##
## See https://uptrace.dev/get/alerting.html for details.
##
alerting:
rules:
- name: Network errors
metrics:
- system.network.errors as $net_errors
query:
- $net_errors > 0 group by host.name
# for the last 5 minutes
for: 5m
annotations:
summary: '{{ $labels.host_name }} has high number of net errors: {{ $values.net_errors }}'
- name: Filesystem usage >= 90%
metrics:
- system.filesystem.usage as $fs_usage
query:
- group by host.name
- group by device
- where device !~ "loop"
- $fs_usage{state="used"} / $fs_usage >= 0.9
for: 5m
annotations:
summary: '{{ $labels.host_name }} has high FS usage: {{ $values.fs_usage }}'
- name: Uptrace is dropping spans
metrics:
- uptrace.projects.spans as $spans
query:
- $spans{type=dropped} > 0
for: 1m
annotations:
summary: 'Uptrace has dropped {{ $values.spans }} spans'
- name: Always firing (for fun and testing)
metrics:
- process.runtime.go.goroutines as $goroutines
query:
- $goroutines >= 0 group by host.name
for: 1m
annotations:
summary: '{{ $labels.host_name }} has high number of goroutines: {{ $values.goroutines }}'
# Create alerts from error logs and span events.
create_alerts_from_spans:
enabled: true
labels:
alert_kind: error
##
## AlertManager client configuration.
## See https://uptrace.dev/get/alerting.html for details.
##
## Note that this is NOT an AlertManager config and you need to configure AlertManager separately.
## See https://prometheus.io/docs/alerting/latest/configuration/ for details.
##
alertmanager_client:
# AlertManager API endpoints that Uptrace uses to manage alerts.
urls:
- 'http://alertmanager:9093/api/v2/alerts'
##
## Various options to tweak ClickHouse schema.
## For changes to take effect, you need reset the ClickHouse database with `ch reset`.
##
ch_schema:
# Compression codec, for example, LZ4, ZSTD(3), or Default.
compression: ZSTD(3)
# Whether to use ReplicatedMergeTree instead of MergeTree.
replicated: false
# Cluster name for Distributed tables and ON CLUSTER clause.
#cluster: uptrace1
spans:
storage_policy: 'default'
# Delete spans data after 30 days.
ttl_delete: 30 DAY
metrics:
storage_policy: 'default'
# Delete metrics data after 90 days.
ttl_delete: 90 DAY
##
## Addresses on which Uptrace receives gRPC and HTTP requests.
##
listen:
# OTLP/gRPC API.
grpc:
addr: ':14317'
# tls:
# cert_file: config/tls/uptrace.crt
# key_file: config/tls/uptrace.key
# OTLP/HTTP API and Uptrace API with UI.
http:
addr: ':14318'
# tls:
# cert_file: config/tls/uptrace.crt
# key_file: config/tls/uptrace.key
##
## Various options for Uptrace UI.
##
site:
# Overrides public URL for Vue-powered UI in case you put Uptrace behind a proxy.
#addr: 'https://uptrace.mydomain.com'
##
## Spans processing options.
##
spans:
# The size of the Go chan used to buffer incoming spans.
# If the buffer is full, Uptrace starts to drop spans.
#buffer_size: 100000
# The number of spans to insert in a single query.
#batch_size: 10000
##
## Metrics processing options.
##
metrics:
# List of attributes to drop for being noisy.
drop_attrs:
- telemetry.sdk.language
- telemetry.sdk.name
- telemetry.sdk.version
# The size of the Go chan used to buffer incoming measures.
# If the buffer is full, Uptrace starts to drop measures.
#buffer_size: 100000
# The number of measures to insert in a single query.
#batch_size: 10000
##
## SQLite/PostgreSQL db that is used to store metadata such us metric names, dashboards, alerts,
## and so on.
##
db:
# Either sqlite or postgres.
driver: sqlite
# Database connection string.
#
# Uptrace automatically creates SQLite database file in the current working directory.
# Make sure the directory is writable by Uptrace process.
dsn: 'file:uptrace.sqlite3?_pragma=foreign_keys(1)&_pragma=busy_timeout(1000)'
# Secret key that is used to sign JWT tokens etc.
secret_key: 102c1a557c314fc28198acd017960843
# Enable to log HTTP requests and database queries.
debug: false