feat: add support for monitor-related flags in provider helm chart (#296

)

- Introduced `monitor` block in `values.yaml` for better organization of
monitoring settings.
- Added support for the following new flags in `statefulset.yaml`:
  - `AKASH_MONITOR_MAX_RETRIES`
  - `AKASH_MONITOR_RETRY_PERIOD`
  - `AKASH_MONITOR_RETRY_PERIOD_JITTER`
  - `AKASH_MONITOR_HEALTHCHECK_PERIOD`
  - `AKASH_MONITOR_HEALTHCHECK_PERIOD_JITTER`
- Updated comments in `values.yaml` for clarity and documentation.


## Refs


akash-network/provider@99cb9ac
#289

charts/akash-provider/Chart.yaml

@@ -17,7 +17,7 @@ type: application
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
 
 # Major version bit highlights the mainnet release (e.g. mainnet4 = 4.x.x, mainnet5 = 5.x.x, ...)
-version: 11.1.0
+version: 11.1.1
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to

charts/akash-provider/templates/statefulset.yaml

@@ -249,6 +249,16 @@ spec:
               value: "{{ .Values.minimumbalance }}"
             - name: AKASH_BID_DEPOSIT
               value: "{{ .Values.bidmindeposit }}"
+            - name: AKASH_MONITOR_MAX_RETRIES
+              value: {{ if (hasKey (.Values.monitor | default dict) "maxRetries") }}{{ (.Values.monitor | default dict).maxRetries | quote }}{{ else }}"40"{{ end }}
+            - name: AKASH_MONITOR_RETRY_PERIOD
+              value: {{ (.Values.monitor | default dict).retryPeriod | default "4s" | quote }}
+            - name: AKASH_MONITOR_RETRY_PERIOD_JITTER
+              value: {{ (.Values.monitor | default dict).retryPeriodJitter | default "15s" | quote }}
+            - name: AKASH_MONITOR_HEALTHCHECK_PERIOD
+              value: {{ (.Values.monitor | default dict).healthcheckPeriod | default "10s" | quote }}
+            - name: AKASH_MONITOR_HEALTHCHECK_PERIOD_JITTER
+              value: {{ (.Values.monitor | default dict).healthcheckPeriodJitter | default "5s" | quote }}
 
           ports:
             - name: api

charts/akash-provider/values.yaml

@@ -93,6 +93,37 @@ ipoperator: false
 
 debug: "false"
 
+#  monitor
+##
+
+# These monitor settings control the retry behavior for deployment checks and health monitoring.
+#
+# If a user deployment cannot start for any reason - such as an issue with the deployment itself,
+# a provider issue (e.g., a lost node due to disk or networking failure), or scheduled maintenance -
+# this monitor starts a counter for retry attempts. It determines how many times and how frequently
+# the provider will retry before marking the deployment as failed and closing the lease.
+#
+# With the default settings (maxRetries = 40, retryPeriod = 4s, retryPeriodJitter = 15s), the total retry time is ~8 minutes.
+#
+# Setting maxRetries to 0 will immediately deem the deployment as failed without attempting any retries.
+#
+# Formula: Total Time = maxRetries * (retryPeriod + average retryPeriodJitter)
+# Example for 1 hour (3600 seconds): maxRetries = 3600 / (retryPeriod + average retryPeriodJitter)
+# Example for 24 hours (86400 seconds): maxRetries = 86400 / (retryPeriod + average retryPeriodJitter)
+# Example for 48 hours (172800 seconds): maxRetries = 172800 / (retryPeriod + average retryPeriodJitter)
+
+# Defaults: retryPeriod = 4s, retryPeriodJitter = 15s, average retryPeriodJitter = 7.5s
+# Example for 1 hour: maxRetries = 3600 / (4 + 7.5) = 3600 / 11.5 ≈ 313
+# Example for 24 hours: maxRetries = 86400 / (4 + 7.5) = 86400 / 11.5 ≈ 7513
+# Example for 48 hours: maxRetries = 172800 / (4 + 7.5) = 172800 / 11.5 ≈ 15026
+
+monitor:
+  maxRetries: 40                     # Maximum retry attempts before closing a lease
+  retryPeriod: 4s                    # Time interval between retries
+  retryPeriodJitter: 15s             # Jitter for retry period
+  healthcheckPeriod: 10s             # Health check period
+  healthcheckPeriodJitter: 5s        # Jitter for health check period
+
 # Percentage of CPU overcommit
 overcommit_pct_cpu: 0
 # Percentage of memory overcommit