Releases: clearlinux/telemetrics-client
v2.1.1
This release contains:
-
Fix the bug where journal probe and klogscanner probe do not restart when telemctl restart command is executed.
-
Fix for potential segmentation fault in telem-record-gen when opt-out is enabled.
-
New probe that reports python exceptions.
-
Reduced telempostd and telemprobd timeouts to 5 minutes to reduce the time these processes are idle. This reduces telemetrics-client resources footprint.
-
crashprobe no longer keeps coredump files under /tmp.
-
hprobe has new arguments to provide locale, installed bundles, and uptime info in payload.
-
journalprobe has a new argument to control journal scanning. Error events logged in the journal while telemetry is disabled (user opted out) are never reported.
-
Merge klogscanner and oopsprobe functionality.
-
Multiple code improvements for correctness and consistency to hello.c, bert_probe.c.
-
Unified implementation and handling of telemetry configuration file override for all probes.
-
Update and rebuild documentation to reflect daemon functionality split.
v2.0.1
v2.0.0
This release contains:
- Refactored telemetry daemon (formerly telemd) from a daemon that handles backend and probes communication to a daemon that is in change of communication with probes only (telemprobd).
- A second daemon (telempostd) was implemented. This daemon transfers telemetry to backend, handles the spool, and applies policies to telemetry messages.
- telem-record-gen has two additional options echo (
-o
or--echo
) and no post (-n
or--no-post
). When using flag echo the telemetry message will be printed to stdout and when using no post option telem-record-gen will not deliver the message for transfer. - crashprobe retry when processing errors occur.
- crash dump configuration removal was reverted from previous release to allow telemetry client to work on other distributions. Any disabling of this configuration should be done during packaging for specific distribution.
v1.16.3
v1.16.2
v1.16.1
v1.16.0
This release contains:
-
record_retention_enabled configuration key: when configuration value
is enabled copies of the last 100 telemetry record payload are kept
locally. -
record_server_delivery_enabled configuration key: when this key is
set to true (default value) records are reported to server. When key
is set to false records are not sent. This configuration can be used
with record_retention_enabled to retain debug information locally
only. -
A new switch (-i/--include_record) for 'telemctl journal' command,
this switch allows the inclusion of record contents when printing
journal to standard output.
v1.15.1
v1.15.0
This release contains:
-
A new header event_id, this value can be used by a probe or probes to
group multiple records when these records are the result of a unique
event in the system, i.e. an update could create multiple records and
these records could be tagged with a single event id. -
Telemetry record logging, this change adds a log of records reported
to telemetry server. The log can be accessed using telemctl journal
sub-command.
v1.14.3
Release contains:
-
Fixes an issue where if the telemetry client bundle is added after
installation required directories were not created until the OS is
restarted. This change makes sure that all the required directories
are created before the telemd service is started. -
Adds an option to telemctl to indicate if telemd.service is active.
-
Removes telemetry client temporary directories on telemetry opt-out.
-
Add a base64 encoding routine to the nica library, this library can
be used by probes to encode binaries for http transport to telemetry
server. -
Fixes configuration parameter on hprobe.
-
Adds a probe to report Boot Error Region Table if this record exists
for more information about BERT, consult ACPI spec