Grafana Data Source
Parseable data source plugin lets you query and visualize logs, metrics, and traces stored in Parseable from your Grafana dashboards, Explore views, and alert rules.
Pre-requisites
- Parseable server running and receiving telemetry from your application.
- Grafana
>= 10.0.0installed and running.
Installation
Several ways to install the plugin:
-
Grafana UI: install from your Grafana instance (Configuration > Data sources > Add Data source > search "Parseable").
-
Grafana CLI:
grafana-cli plugins install parseable-parseable-datasource. Restart Grafana. -
Grafana Helm Chart: add under the
pluginssection ofvalues.yaml:
plugins:
- https://grafana.com/api/plugins/parseable-parseable-datasource/versions/2.0.0/download;parseable-parseable-datasourceAdd under the datasources section:
## ref: http://docs.grafana.org/administration/provisioning/#datasources
datasources:
datasources.yaml:
apiVersion: 1
datasources:
- name: Parseable
type: parseable-parseable-datasource
url: http://parseable.parseable
access: proxy
isDefault: true
basicAuth: true
basicAuthUser: admin
secureJsonData:
basicAuthPassword: adminUpdate url, basicAuthUser, and basicAuthPassword for your instance.
To install the plugin alongside Grafana itself, set GF_INSTALL_PLUGINS=parseable-parseable-datasource 2.0.0 and restart Grafana.
For Ansible / config-management provisioning, see the Grafana provisioning docs.
Configuration
Set the URL and port of your Parseable server (e.g. https://<your-parseable-host>:8000). Parseable uses basic auth — toggle Basic Auth under the Auth section and enter the credentials under Basic Auth Details.
Click Save & Test. If it fails, check credentials and connectivity.
Query Modes
The plugin auto-selects available editor modes based on the dataset type of the selected stream.
Logs and Traces — Builder or SQL
- Builder: dropdown-driven UI for filters, aggregates, time range, and ordering. Filter values populate from
dataset_statswith top-5 suggestions and an inline Show more for the full set. - SQL (Code): full SQL editor with autocomplete for fields, types, and operators. Double-quote wrapping for identifiers happens automatically.
Metrics — Builder or PromQL
Metrics streams expose Prometheus-compatible endpoints.
- Builder: Prometheus-style UI — pick a metric name, add label matchers (
=,!=,=~,!~), live PromQL preview of the assembled selector. Label values are scoped to the selected metric and time range. - PromQL (Code): full PromQL editor with Lezer-based syntax highlighting, completion (metric names, labels, functions), signature help, hover docs, and inline error markers.
Switching between modes preserves your selection where possible (Builder selectors compose into a valid PromQL selector when you flip to Code).
Alerts
Grafana alerts work against any of the three dataset types.
- Logs / traces: use Builder or SQL to produce numeric reducers (
count,avg,sum, etc.). - Metrics: use PromQL with the standard
rate(...),sum by (...), threshold patterns.
Setting Up an Alert
- Navigate to
your-domain:port/alerting/new/alerting. - Add a unique rule name.
- Add a query that returns numeric values.
- Specify threshold behavior for the defined rule query.
- Define evaluation behavior (frequency, grouping).
- Save the alert.
Managing Alerts
List, edit, or delete alerts at your-domain:port/alerting/list.
Notification Channels
Configure receivers at your-domain:port/alerting/notifications/receivers/new. Grafana supports Alertmanager, Email, Slack, Telegram, Webhook, Microsoft Teams, and more.
Notification Policies
Route alerts to the right channels at your-domain:port/alerting/routes.
The alerts list page shows current status, last fired/evaluated times, and the rule query results.
Mapping Non-Numeric Values for Alerts
This applies to SQL queries on logs / traces. PromQL queries on metrics don't need this — metric values are already numeric, label matchers filter strings, and comparison operators return 0/1 directly.
Grafana alert evaluators compare numeric values. Map string statuses to numbers in SQL when alerting on log fields:
SELECT
...
CASE
WHEN status = 'OK' THEN 0
WHEN status = 'WARN' THEN 1
WHEN status = 'ERROR' THEN 2
END AS status_value
FROM
table_nameWas this page helpful?