Fetching contributors…
Cannot retrieve contributors at this time
197 lines (120 sloc) 4.22 KB
Fix bug in skipping bad input lines to karbon.
Upgrade to kairos 0.10.0 with breaking change for custom transforms.
Switch to python logger from print statements.
Switched to `logging.exception()` from `traceback.print_exc()` calls.
Added on_load and on_reload callbacks for configurations.
Added `schema_test` tool for performance testing.
Uses kairos 0.9.2 with URL support.
Changed "host_settings" to "client_config" as they're passed directly to kairos.
Fix for single character value regression introduced in 0.6.2 fix for issue 10.
If debugging, print torus web requests.
On configuration reload, reset the debug state. Fixes
Fixed matching against stat names with spaces in the name. Fixes
Added "/insert" URI for inserting data with HTTP.
Fix the incorrect assignment operator in "/properties".
Added quick start guide and supporting configurations.
Added support for SQL and Cassandra databases in schemas.
Added "bin/migrate" tool for migrating data between schemas.
Default database is in-memory "sqlite:///:memory:"
Added "host_settings" schema option.
Removed support for "/data" URI.
Added "/list" and "/properties" URIs.
Fixed significant performance degradation in all endpoints by correctly
using WSGI.
Renamed "condensed" parameter in "/series" to "condense" to match
latest kairos.
Added support for macros, which are the basis for exposing kairos' custom
read functionality.
Fixed matching when there's a list of expressions and a list of stat names.
Include command line arguments when changing process title
Adds "rolling" option to schemas, allowing data to be inserted into multiple
intervals at a time. This increases storage space but supports rapid
calculation of aggregates over pre-determined interval ranges.
Stat queries can be comma-separated values to join multiple timeseries.
Added "collapse" parameter to "/series" which supports collapsing all of the
data in the date range into a single result. This supports easy computation of
aggregates across a time range, such as "maximum value in the last 7 days".
Schemas can include "TRANSFORMS" stanza for defining a named map of data
transforms that will be used when querying torus.
Adding hiredis requirement as redis is not performant enough without it.
Rename torus and karbon processes for easy "kill -SIGHUP `pidof karbon`".
Configuration files can include "DEBUG" stanza for turning on debug logging.
Debugging output currently only implemented in karbon. See the "Debugging"
section in the README for details.
Renamed command line argument "--schema" to "--config" to better reflect the
nature of the files.
Result of "/series" now includes the full stat name and its function in the
'stat' field. Added 'stat_name' field in results.
Fixed handling of "start" and "end" parameters to "/series"
Fixed handling of "condensed" and "interval" parameters to "/series"
Upgrade to kairos 0.3.0, adds support for Gregorian dates and date ranges
to "/series"
Adds support for UNIX timestamps or parsedatetime-compatible strings for
'start' and 'end' parameters to "/series"
Adds support for 'steps' parameter to "/series"
Fix bug in detecting recursion when processing aggregates
Added support for Mongo backends via latest kairos
Require the database type in the URI
Redis and Mongo support full suite of URI options
Added "schema_debug" binary to package
Aggregates can now cascade to other aggregates, with a simple failsafe on
infinite loops
Update to kairos 0.1.5 with a bug fix for reading data from timeseries with
relative time interval schemas.
Added support for transforms in a schema, to replace or suppress key/value
Upgrade to kairos 0.1.4 with bug fixes and performance improvements, relative
date support and gauge type
Fixed redis host parsing when no port defined
Change command line parameter "--config" to "--schema"
Initial release of torus and karbon