Releases: TykTechnologies/tyk-pump
v1.3.0
Changelog
General improvements
- Multiple analytics keys - better key purge distribution across Redis cluster (#320)
- Improved logs (#319)
- Revert of “obfuscate keys” feature (#315)
- All the pumps can be configurable by environment variables (#333)
Pumps improvements
- Splunk Pump - Customizable analytics fields (#317)
- Splunk Pump - Bugfix when using ssl_insecure_skip_verify:false and no certs(#336).
- CSV Pump - bugfix when writing some analytics records values (#329).
Docker images
docker pull tykio/tyk-pump-docker-pub:slim
docker pull tykio/tyk-pump-docker-pub:s1.3
docker pull docker.cloudsmith.io/tyk/tyk-pump/tyk-pump:slim
docker pull docker.cloudsmith.io/tyk/tyk-pump/tyk-pump:v1.3
docker pull tykio/tyk-pump-docker-pub:v1.3.0
docker pull tykio/tyk-pump-docker-pub:v1.3
docker pull docker.cloudsmith.io/tyk/tyk-pump/tyk-pump:v1.3.0
docker pull docker.cloudsmith.io/tyk/tyk-pump/tyk-pump:v1.3
v1.2.0
Improving our third party integrations
Moesif - @keyur9 @dkm199 Added new config options including EventQueueSize, BatchSize, and TimerWakeUpSeconds
as well as support to fetch userId from authorization header. #291
Datadog - @diegobernardes added an option to configure the tags pushed to DogStatsD so that it reduces the cost of running datadog. #292
Prometheus - @Bouwdie fixed a bug related to the keystatus counter. #295
Improving Pump resiliency
Introduce purge configuration to avoid data loss, improve load distribution when multiple pumps are set up. #203
Ability to reduce data load by removing automatic aggregations. #301
Ability to configure mongo pump timeout so that you can avoid pump freezing if mongo stops #302
Other Features / Bugs
Logzio HTTP proxy support #294 - and the bug it generated in Kafka auth #308
Ability to obfuscate keys for all records #309
Blurring sensitive data coming from mongo #285
v1.1.0
New features and enhancements:
- Implemented syslog pump (#189).
- Added
omit_detailed_recording
setting (#269). - Added key obfuscation for Splunk pump (#140).
- Added mutual TLS support for Kafka pump (#235).
- Added basic auth support for Kafka and ElasticSearch pumps (#247).
- Upgraded Moesif Pump dependency. Added support for request and response logging, events sampling and field masking (#263).
Bug fixes:
v1.0.1
v1.0.0
v0.8.5.1
v0.8.5
v0.8.4
0.8.3
-
Handled unsupported MongoDB characters [#113]
MongoDB doesn't support.
in parent field names. Should a path contain.
, mongoDB would get corrupted. We have now replaced the.
character with its Unicode equivalent. -
Added support for pumping to an Elasticsearch Cluster [#180]
Previously, it was only possible to pump analytics to a single Elasticsearch endpoint. Due to this limitation, it was not previously possible to support Elasticsearch clusters. -
Resolved issue where aggregate TCP proxy analytics were incorrectly calculated [#182]
0.8.2
- Added way to ignore tags to be aggregated, which can cause aggregated records to go too big (if you have lot of unique tags, like request-id). Added new option to Aggregate and Hybrid pumps:
ignore_tag_prefix_list
, which should be a string array with tag prefixes you want to exclude. Additionally, it now tries to detect such scenarios and print a warning with list of tags that can cause issues. You can control the warning threshold using thethreshold_len_tag_list
parameter. #162 - Mongo aggregate pump now can store data in minute chunks instead of hourly ones, if you set
store_analytics_per_minute
totrue
to aggregate or hybrid pump. This can help with issues when you have lot of keys or tags, generating record of size more than 16 megabytes. #164 - Mongo pump, instead of skipping records which out of the size, should truncate them instead #154
- Added new attributes to ElasticSearch pump:
RawPath
,ContentLength
, andTags
#157 - ElasticSearch pump now can generate "id" attribute in order to ensure record uniqueness. Set
generate_id
inside ES pump config to enable new behavior. #169 - "debug" logging should print only the number of records instead of their content #159