2.2.18 "Zack" - Nov 5, 2024

Changes:

• Introducing a a new config key agent_status_timeout and setting it to 30s by default. Until version 2.2.17 it was hardcoded to 5 seconds.

Fixes:

• Fixed a bug where pod logs were removed from ingestion because of a temporrary K8s API error and the agent was not able to add them again.

2.2.17 "Sinthia" - Sep 17, 2024

Changes:

• Several packages were upgraded to mitigate vulnerabilities - orjson==3.10.7, PyMySQL==1.1.1
• Unused root CA file scalyr_agent_ca_root was removed
• Alpine base image upgraded to 2.19.4

Fixes:

• Added safety check when handling pending addEvents task to handle the following error: AttributeError: 'AddEventsTask' object has no attribute '_CopyingManagerWorkerSession__receive_response_status'
• Added missing zstandard library to alpine image

2.2.16 "Lenny" - May 29, 2024

Changes:

• Fixed steady growth of memory usage of Syslog Monitor. syslog_processing_thread_count can be used to limit the number of threads and memory consumption.
• Added the max_allowed_checkpoint_age option to allow configuring the checkpoint age (default 15 minutes), useful for agents monitoring a large number of short-lived logs
• Added the Kubernetes monitor k8s_cri_query_filesystem_retain_not_found option to disable caching info about pods that are dead (default is true), useful for agents that monitor clusters whose pods are not cleaned up frequently

2.2.15 "Jain" - May 20, 2024

Features:
Ability to ignore k8s labels (#1213).
A user can now define glob filters for including and excluding container and controller labels in logs.
For label to be included it needs to match one of the globs in k8s_label_include_globs and none of the globs in k8s_label_exclude_globs.

From documentation:
| k8s_label_include_globs | Optional, (defaults to ['*']). Specifies a list of K8s labels to be added to logs. |
| k8s_label_exclude_globs | Optional, (defaults to []]). Specifies a list of K8s labels to be ignored and not added to logs. |

Follows the same logic as label_include_globs and label_exclude_globs from the docker_monitor.

2.2.14 "Tycho" - May 6, 2024

Changes:
Support sending events to multiple Dataset accounts via K8s annotations. This feature was updated to use secrets from the same namespace as the pod and hierarchy of default secrets were added: default secret -> namespace secret -> pod secret -> container secret.
See kubernetes_monitor.md for documentation.

2.2.13 "Killian" - Mar 30, 2024

Feature:

• Kubernetes Monitor - when K8s Pod API returnes 404, the logs are ingested based on SCALYR_K8S_INCLUDE_ALL_CONTAINERS regardless of pods annotations. Short-lived pods are usually affected.

2.2.12 "Gizmo" - Feb 29, 2024

Feature:

• Support sending events to multiple Dataset accounts via K8s annotations. See kubernetes_monitor.md for documentation.

2.2.11 "Maxson" - Feb 7, 2024

Security fix:

• Alpine image upgraded to 3.19.1 to mitigate CVE-2023-7104 and CVE-2023-5363

2.2.10 "Aradesh" - Feb 5, 2024

Improvement:

• Windows Event Log Monitor - EventID field has been expanded into three fields. Until now EventID from the Windows Event Log was represented only by EventID field computed the following way win32api.MAKELONG(EventID, EventIDQualifiers). EventID and EventIDQualifiers now hold the values from the Windows Event Log without any modification and InstanceID is computed from EventID and EventIDQualifiers as described above.

Fix:

• Missing monitor field added to the event data sent by Syslog Monitor. The following field was added - ‘monitor’:'agentSyslog'.

2.2.9 "Tandi" - Jan 11, 2024

2.2.9 "Tandi" - Jan 11, 2024

Feature:

• Support for CRI multi lines - implementing support for the following format: https://github.com/kubernetes/design-proposals-archive/blob/main/node/kubelet-cri-logging.md

Fix:

• Python2 suport for Syslog monitor fixed

Security Fix:

• Alpine image upgraded from 3.18.13 to 3.18.14 to fix recently found vulnerabilities