[TOOLS-1740] - Export all histogram buckets by default for latency me…

…trics (#46)

- Exports all histogram buckets by default
- Adds new configuration 'latency_buckets_count' to control number of buckets to be exported

README.md

@@ -124,6 +124,14 @@ Packages will be generated under `./pkg/target/` directory.
     db_host="localhost"
     db_port=3000
     ```
+
+- Configure timeout (in seconds) for info commands to Aerospike node (optional). Default value is `5` seconds.
+    ```toml
+    [Aerospike]
+    # timeout for sending commands to the server node in seconds
+    timeout=5
+    ```
+
 - Update Aerospike security and TLS configurations (optional),
     ```toml
     [Aerospike]
@@ -302,4 +310,25 @@ Packages will be generated under `./pkg/target/` directory.
     "admin",
     "superuser"
     ]
-    ```
+    ```
+
+- Exporter logs to console by default. To enable file logging, use `log_file` configuration to specify a file path. Use `log_level` configuration to specify a logging level (optional). Default logging level is `info`.
+    ```toml
+    [Agent]
+    # Exporter logging configuration
+    # Log file path (optional, logs to console by default)
+    # Level can be info|warning,warn|error,err|debug|trace ('info' by default)
+    log_file = ""
+    log_level = ""
+    ```
+
+- Use `latency_buckets_count` to specify number of histogram buckets to be exported for latency metrics (optional). Bucket thresholds range from 2<sup>0</sup> to 2<sup>16</sup> (`17` buckets). All threshold buckets are exported by default (`latency_buckets_count=0`).
+
+    Example, `latency_buckets_count=5` will export first five buckets i.e. `<=1ms`, `<=2ms`, `<=4ms`, `<=8ms` and `<=16ms`.
+    ```toml
+    [Aerospike]
+    # Number of histogram buckets to export for latency metrics. Bucket thresholds range from 2^0 to 2^16 (17 buckets).
+    # e.g. latency_buckets_count=5 will export first five buckets i.e. <=1ms, <=2ms, <=4ms, <=8ms and <=16ms.
+    # Default: 0 (export all threshold buckets).
+    latency_buckets_count=0
+    ```

ape.toml

@@ -102,6 +102,11 @@ auth_mode=""
 # timeout for sending commands to the server node in seconds
 timeout=5
 
+# Number of histogram buckets to export for latency metrics. Bucket thresholds range from 2^0 to 2^16 (17 buckets).
+# e.g. latency_buckets_count=5 will export first five buckets i.e. <=1ms, <=2ms, <=4ms, <=8ms and <=16ms.
+# Default: 0 (export all threshold buckets).
+latency_buckets_count=0
+
 # Metrics Allowlist - If specified, only these metrics will be scraped. An empty list will exclude all metrics.
 # Commenting out the below allowlist configs will disable metrics filtering (i.e. all metrics will be scraped).
 

ape.toml.template

@@ -102,6 +102,11 @@ auth_mode="${AS_AUTH_MODE}"
 # timeout for sending commands to the server node in seconds
 timeout=${TICKER_TIMEOUT}
 
+# Number of histogram buckets to export for latency metrics. Bucket thresholds range from 2^0 to 2^16 (17 buckets).
+# e.g. latency_buckets_count=5 will export first five buckets i.e. <=1ms, <=2ms, <=4ms, <=8ms and <=16ms.
+# Default: 0 (export all threshold buckets).
+latency_buckets_count=${LATENCY_BUCKETS_COUNT}
+
 # Metrics Allowlist - If specified, only these metrics will be scraped. An empty list will exclude all metrics.
 # Commenting out the below allowlist configs will disable metrics filtering (i.e. all metrics will be scraped).
 

config.go

@@ -47,6 +47,8 @@ type Config struct {
 
 		Timeout uint8 `toml:"timeout"`
 
+		LatencyBucketsCount uint8 `toml:"latency_buckets_count"`
+
 		UserMetricsUsersAllowlist []string `toml:"user_metrics_users_allowlist"`
 		UserMetricsUsersBlocklist []string `toml:"user_metrics_users_blocklist"`
 

docker-entrypoint.sh

@@ -24,6 +24,7 @@ export AS_AUTH_MODE=${AS_AUTH_MODE:-""}
 export AS_AUTH_USER=${AS_AUTH_USER:-""}
 export AS_AUTH_PASSWORD=${AS_AUTH_PASSWORD:-""}
 export TICKER_TIMEOUT=${TICKER_TIMEOUT:-5}
+export LATENCY_BUCKETS_COUNT=${LATENCY_BUCKETS_COUNT:-0}
 export NAMESPACE_METRICS_ALLOWLIST=${NAMESPACE_METRICS_ALLOWLIST:-""}
 export SET_METRICS_ALLOWLIST=${SET_METRICS_ALLOWLIST:-""}
 export NODE_METRICS_ALLOWLIST=${NODE_METRICS_ALLOWLIST:-""}

latency_parser.go

@@ -15,7 +15,7 @@ import (
 // error-no-data-yet-or-back-too-small;
 // or,
 // {test}-write:;
-func parseLatencyInfoLegacy(s string) map[string]StatsMap {
+func parseLatencyInfoLegacy(s string, latencyBucketsCount int) map[string]StatsMap {
 	ip := NewInfoParser(s)
 	res := map[string]StatsMap{}
 
@@ -103,8 +103,8 @@ func parseLatencyInfoLegacy(s string) map[string]StatsMap {
 			bucketLabels = append(bucketLabels, strings.Trim(bLabels[i], "><=ms"))
 			bucketValuesFloat = append(bucketValuesFloat, v)
 
-			// Stop after first zero bucket. No point processing further buckets.
-			if val == 0 {
+			if latencyBucketsCount > 0 && i >= latencyBucketsCount {
+				// latency buckets count limit reached
 				break
 			}
 		}
@@ -138,7 +138,7 @@ func parseLatencyInfoLegacy(s string) map[string]StatsMap {
 // Format (with and without latency data)
 // {test}-write:msec,4234.9,28.75,7.40,1.63,0.26,0.03,0.00,0.00,0.00,0.00,0.00,0.00,0.00,0.00,0.00,0.00,0.00,0.00;
 // {test}-read:;
-func parseLatencyInfo(s string) map[string]StatsMap {
+func parseLatencyInfo(s string, latencyBucketsCount int) map[string]StatsMap {
 	ip := NewInfoParser(s)
 	res := map[string]StatsMap{}
 
@@ -215,8 +215,8 @@ func parseLatencyInfo(s string) map[string]StatsMap {
 			bucketLabels = append(bucketLabels, strconv.Itoa(1<<(i-1)))
 			bucketValuesFloat = append(bucketValuesFloat, v)
 
-			// Stop after first zero bucket. No point processing further buckets.
-			if val == 0 {
+			if latencyBucketsCount > 0 && i >= latencyBucketsCount {
+				// latency buckets count limit reached
 				break
 			}
 		}

watcher_latency.go

@@ -35,9 +35,9 @@ func (lw *LatencyWatcher) refresh(o *Observer, infoKeys []string, rawMetrics map
 	var latencyStats map[string]StatsMap
 
 	if rawMetrics["latencies:"] != "" {
-		latencyStats = parseLatencyInfo(rawMetrics["latencies:"])
+		latencyStats = parseLatencyInfo(rawMetrics["latencies:"], int(config.Aerospike.LatencyBucketsCount))
 	} else {
-		latencyStats = parseLatencyInfoLegacy(rawMetrics["latency:"])
+		latencyStats = parseLatencyInfoLegacy(rawMetrics["latency:"], int(config.Aerospike.LatencyBucketsCount))
 	}
 
 	log.Tracef("latency-stats:%+v", latencyStats)