OpenObserve
https://openobserve.ai/OpenObserve is an open-source observability platform for logs, metrics and traces, queried with SQL.
Installation
Required packages: sqlalchemy-openobserve
Connection String
openobserve://{email}:{password}@{host}:{port}/{organization}Connection Examples
Local server over plain HTTP
openobserve://root%40example.com:password@localhost:5080/default?ssl=falseSupported Features
JOINsSubqueriesDynamic SchemaCatalog SupportDynamic CatalogSSH TunnelingQuery CancellationFile UploadUser ImpersonationCost EstimationSQL Validation
Feature Score: 45/201
Time Grains
Common Time Grains:
SECONDMINUTEHOURDAYWEEKMONTHQUARTERYEAR
Extended Time Grains:
FIVE_SECONDSTHIRTY_SECONDSFIVE_MINUTESTEN_MINUTESFIFTEEN_MINUTESTHIRTY_MINUTESHALF_HOURSIX_HOURSWEEK_STARTING_SUNDAYWEEK_STARTING_MONDAYWEEK_ENDING_SATURDAYWEEK_ENDING_SUNDAYQUARTER_YEAR
Extended Aggregate Functions
The base aggregates (SUM, COUNT, AVG, MIN, MAX, COUNT DISTINCT) are supported everywhere; these are opt-in per engine.
MEDIANSTDDEV_SAMPVAR_SAMP
Troubleshooting
Common error messages you may encounter when connecting to or querying OpenObserve, along with their causes and solutions.
Notes
The username is your OpenObserve email. Write its @ as %40, and percent-encode special characters in the password. HTTPS is used by default; add ?ssl=false for a plain HTTP server. Each log stream is a table, and its _timestamp column is set up as the time column. Queries are read-only; SHOW and SELECT 1 are not supported by the server. Known server limits: a SELECT with ORDER BY and LIMIT that has no aggregate can return fewer rows than the limit, often none, and SELECT * over a CTE or subquery returns no columns when it matches no rows.
Resources