Telegraf Configuration
- 1 Kinesis Input Consumer
- Several Processors (24 totals) : Renaming / Pivot / Starlark
- influxdb_v2 Output
When deploying my Telegraf Application on an AWS ECS Task the Memory Utilization starts growing and never stops, this indicates that I have a memory leakage somewhere.
Agent
[agent]
interval = "5s"
flush_interval = "5s"
metric_batch_size = 2000
metric_buffer_limit = 10000
skip_processors_after_aggregators = false
debug = true
Input
[[inputs.kinesis_consumer]]
region = "us-east-1"
streamname = "telemetry"
shard_iterator_type = "LATEST"
data_format = "json_v2"
[[inputs.kinesis_consumer.json_v2]]
measurement_name_path = "topic.deviceId"
timestamp_path = "payload.timestamp"
timestamp_format = "unix_ms"
[[inputs.kinesis_consumer.json_v2.tag]]
path = "topic.groupId"
rename = "group_id"
[[inputs.kinesis_consumer.json_v2.tag]]
path = "topic.edgeNodeId"
rename = "project_id"
[[inputs.kinesis_consumer.json_v2.tag]]
path = "topic.deviceId"
rename = "device_id"
[[inputs.kinesis_consumer.json_v2.object]]
path = "payload.metrics"
disable_prepend_keys = true
included_keys = ["value"]
tags = ["name"]
Memory Profile
Task Logs
What could be the potential cause for this to happen?
