- Upgrade dependencies
- Upgrade dependencies
- Upgrade dependencies
- Fix typings
- Fix typings
- Upgrade dependencies, major new ES client version
- Upgrade dependencies
- Expose option
internalLogger
for last resort logger
- Add more debug output
- Upgrade dependencies
- Fix duplicate entries when buffering is used
- Fix last resort error logging
- Fix last resort error logging
- Upgrade optional APM dependency for making it work with node 16.
- Upgrade ES client.
- Expose transformer function so that it can be adapted.
- Upgrade ES client
- Add console.log output in case of bulk writer errors
- Fix indexPrefix can be of type string | Function
- Fix travis build process (ES JS Client (7.11.2 instead of 11), nodejs 8 deprecated on yargs)
- Fix templateName is undefined or null if options.index is undefined or null
- Correct npm link in README.md (browsenpm.org isn't connected to npm anymore)
- Prevents resending successfully sent items.
- Add
retryLimit
option - Fix index name not being passed to bulk writer
- Fix index name not being passed to bulk writer
- Introduce
source
parameter - Fix some datastream issues
- Remove _type and don't write it
- Rename config Property
ensureMappingTemplate
toensureIndexTemplate
andmappingTemplate
accordingly - Make Index Templates not appear as Legacy ones in Kibana, use
putIndexTemplate
instead ofputTemplate
- ES 11 JS Client
- Support for data streams
- Drop support for ES <= 6
- Increase retries from 5 to 400 (almost 7 minutes) as some massive uServices systems take a while until they spun up
- Inject indexing error in case writing to ES fails
- Use latest ES client
- Docs fixes
- Make retry mechanism give up after 5 retries
- Make emit a warning only when the max retry has been reached
- Make APM param optional
- Move APM dependency into normal dependencies
- Improve typings
- Added flush() method to ES transport
- Emit transport internal errors no longer as
error
but aswarning
because even when listened to with.on('error', ...)
it lead to anUnhandledPromiseRejectionWarning
.
- Upgrade ES Client
- New approach to error handling where, in case of an emitted error the source stream is automatically re-attached to the transport stream
- This commit adds a retry counter - after the number of retries has been exceeded then the document is discarded.
- Exposes healthcheck options
- ES version specific mapping templates
- Fix typings
- Fix typings
- Remove default export from typings
- Upgrade dependencies
- Correct order for entries with same timestamp
- Fix edge case with
this.client.bulk
- Upgrade typings
- Make compatible with ES client 7.6
- Upgrade deps
- Make compatible with ES client 7.5
- Add
-*
to given index pattern prefix for creating index pattern
- Upgrade deps
- Use
existsTemplate
instead ofgetTemplate
- Make buffering in case of outages more stable
- Don't provide
type
anymore with ES client's bulk operation - Default to
all
forwaitForActiveShards
- Emit more
error
events in error cases - Adapt default index template to conform with newer ES version
- Update typings
- Switch to new official ES client
- Emit
error
events when an error happens
- Upgrade to ES client
- Remove unneeded dependency
- Reduce package size by swapping moment for days.js and other measured
- Upgrade to ES client v16
- Allow to pass the timestamp through log() so that it's not generated by the transformer
- Fix 2 severe issues described here: #87
- Prevent .git from being published to npm
- Properly implement non buffering case --> immediately write any message to ES
- Upgrade deps and several other cleanups
- Upgrade of typings
- Correct documentation that we use
_doc
as default type`
- Upgrade deps, also Winston 3.2.1
- Incorporate fix to prevent flush called twice after each interval instead of once
- Better ignore extraneous resources pushed to npm
- Upgrade deps, also Winston 3.2.0
- Update the documentation
- Add feature to limit message buffering
- Upgrade deps
- Allow the process to exit when the logger is ended.
- Upgrade to Winston 3.1.0
- Documentation fixes
- Make sure no messages are lost in case of an ES level fault
- Add a Typescript declaration file
- Fix bug in ES fault handler
- Possibility to have an
indexInterfix
- Remove _all field from default mapping as it is deprecated in ES
- Support
logger.info({ message: 'Test', foo: 'bar' });
signature as well (make sure that fieldfoo
ends up infields
in the index)
- Correct usage docs in README.md
- Clone opts of ES client elastic/elasticsearch-js#33
- Upgrade to winston 3.0 and elasticsearch 15.0
- Actually use message type _doc as default
- Mapping template changed --> if you have a custom mapping please check!
- Remove a lot of unsupported settings in the default mapping. CHECK YOR MAPPING, if you are using a custom one
- Remove default mapping due to the fact that they are deprecated in ES v6.0.0 and greater
- See https://www.elastic.co/guide/en/elasticsearch/reference/current/removal-of-types.html
- Fix bug that lead to mapping template not being used
- Upgrade deps
Better error handling.
- In order to prevent
UnhandledPromiseRejectionWarning
and tackle node.js deprecation DEP0018, catching and logging to console is now the default way of handling internal errors - Enable
sniffOnConnectionFault
on ES client by default - Change default mapping:
template
-->index_patterns
- Migrate default mapping according to https://www.elastic.co/blog/strings-are-dead-long-live-strings
- Moved retry logic into bulkwriter to handle intermittent connectivity interruptions gracefully AND be able to resume operation
- Connection testing is now running infinitely which means the logger never gives up
- Messages logged during a connection outage are buffered
- Rename debug key from
bulk writer
towinston:elasticsearch
- use
finally()
instead ofthen()
to schedule bulk writes even in case of exceptions
- Fix issue with loading built-in mapping
- Upgrade to Elasticsearch client 14 (Elasticsearch 6)
- Ignore 404 errors for silent creation of new indexes
- Upgrade to Winston 2.4.0
- Add pipeline option for elasticsearch
- Upgrade all deps
- Fix linting issues
- Fix loading of template file previously done with require()
- Release for Elasticsearch 5
- Remove
consistency
option - Introduce
waitForActiveShards
option
- Allow
consistency
to be disabled usingfalse
as value - Upgrade deps
- Add timestamp automatically to log messages
- Minimum node.js version 6
- Version upgrades
- Migrate to eslint from jshint and jscs
- Fix for dependencies - move winstom from devDependencies to dependencies in package.json
- Test with ES 2.3.1
- Add time driven bulk write functionality.
- Remove retry functionality for writes as now covered by bulk writer.
- Tests for ES unavailable case.
- Fix for Windows platform -- make default mapping file readable.
- ES 2.1 support (driver update).
- Support ES 2.0, really.
- Support for single index usage, without data suffix via
index
option. - Fix bug when ES client is provided.
- Support ES 2.0.
- Add jshint jscs.
- Minor cleanups.
- Add transport name.
- Call super constructor for inheritance.
- 90% rewrite.
- Use current dependencies.
- Removed feature to generate process stats automatically.
- Removed functionality to generate other implicit fields like @tags.
- Add transformer functionality to modify structure of messages before logging.
- Introduce connection state checking and basic retry mechanism for main ES operations.