diff --git a/.gitignore b/.gitignore
index f6f89d89..94664367 100644
--- a/.gitignore
+++ b/.gitignore
@@ -133,3 +133,4 @@ dmypy.json
 .vscode
 
 node_modules/
+.DS_Store
\ No newline at end of file
diff --git a/prometheus/README.md b/prometheus/README.md
index 804ed856..c1c2d6aa 100644
--- a/prometheus/README.md
+++ b/prometheus/README.md
@@ -115,15 +115,16 @@ services:
       - "./prometheus_config.yml:/etc/prometheus/prometheus.yml" # Assumes prometheus_config.yml exists in your CWD
 ```
 
-## Available Metrics
+## Metrics
 
-The Prometheus extension exposes various LocalStack metrics through the `/_extension/metrics` endpoint, including:
-- Request counts by service
-- Request latencies
-- Resource utilization
-- Error rates
+The Prometheus extension exposes various LocalStack and system metrics through the `/_extension/metrics` endpoint.
 
-For a complete list of available metrics, visit the endpoint directly at `localhost.localstack.cloud:4566/_extension/metrics` when LocalStack is running.
+For a complete list of available metrics, view the:
+- [LocalStack Metrics documentation](./docs/localstack_metrics.md) 
+- [System Metrics documentation](./docs/system_metrics.md) 
+- Otherwise, visit the endpoint directly at `localhost.localstack.cloud:4566/_extension/metrics` when LocalStack is running.
+
+We've also included a [collection of PromQL queries](./docs/event_analysis.md) that are useful for analyzing LocalStack event source mappings performance.
 
 ## Licensing
 
diff --git a/prometheus/docs/event_analysis.md b/prometheus/docs/event_analysis.md
new file mode 100644
index 00000000..17a95058
--- /dev/null
+++ b/prometheus/docs/event_analysis.md
@@ -0,0 +1,104 @@
+# PromQL Queries for Event Processing Statistics
+
+The following queries can be used to analyse performance of LocalStack's event processing capabilties.
+
+## Average Propagation Delay from Event Source to Poller
+
+The average amount of time a record has to wait before being processed during the last 5 minutes. A high propagation delay indicates that our event pollers are taking too long to ingest new events from an event source.
+
+```
+rate(localstack_event_propagation_delay_seconds_sum[5m]) / rate(localstack_event_propagation_delay_seconds_count[5m])
+```
+
+**Example**:
+![Average Propagation Delay](images/avg_propagation_delay.png)
+
+## Batch Efficiency
+
+A ratio showing how efficiently are our pollers retrieving records from an event source relative to how large their maximum batch size is. A higher number indicates that batch sizes could be increased.
+
+```
+rate(localstack_batch_size_efficiency_ratio_sum[1m]) / rate(localstack_batch_size_efficiency_ratio_count[1m])
+```
+
+Example:
+![Batch Efficiency Ratio](images/batch_efficiency_ratio.png)
+
+## Records Per Poll
+
+The average number of records being pulled in by an event poller per minute. When used in conjunction with batch efficiency, you can interpret the performance of your batching configuration.
+
+```
+rate(localstack_records_per_poll_sum[1m]) / rate(localstack_records_per_poll_count[1m])
+```
+
+Example:
+
+![Records Per Poll](images/records_per_poll.png)
+
+## In-Flight Events
+
+Gauges how many events are currently being processed by a target at a given point in time. If event processing is taking long, this is a good way of measuring back-pressure on the system.
+
+```
+localstack_in_flight_events
+```
+
+Example:
+![In-Flight Events](images/in_flight_events.png)
+
+## Event Processing Duration
+
+The average duration per minute that targets are processing events for.
+
+```
+rate(localstack_process_event_duration_seconds_sum[1m]) / rate(localstack_process_event_duration_seconds_count[1m])
+```
+
+Example:
+
+![Event Processing Duration](images/event_processing_duration.png)
+
+## High Latency Event Processing
+
+Retrieve the 95th percentile of processing times in a 5m interval grouped by LocalStack service and operation. Useful for analysing the tail-latency of event processing since this is likely where bottlenecks in performance start to show.
+
+```
+histogram_quantile(0.95, sum by(service, operation, le) (rate(localstack_request_processing_duration_seconds_bucket[5m])))
+```
+
+Example:
+![High Latency Event Processing](images/high_latency_event_processing.png)
+
+## Empty Poll Responses
+
+The approximate number of empty poll requests in a 5 minute interval.
+
+```
+rate(localstack_poll_miss_total[5m]) * 60
+```
+
+Example:
+![Empty Poll Responses](images/empty_poll_responses.png)
+
+## Number of LocalStack requests Processed
+
+The average number of request processed by the LocalStack gateway per minute. This is grouped by service type (i.e SQS) and operation type (i.e ReceiveMessage)
+
+```
+sum by(service, operation) (rate(localstack_request_processing_duration_seconds_count[1m]) * 60)
+```
+
+Example:
+![Requests Processed](images/requests_processed.png)
+
+## In-Flight Requests Against LocalStack Gateway
+
+Measures how many requests the Kinesis, SQS, DynamoDB, and Lambda services are currently processing in a given minute interval. Useful for seeing how hard a given service is currently being hit and the operation type.
+
+```
+sum_over_time(localstack_in_flight_requests{service=~"dynamodb|kinesis|sqs|lambda"}[1m])
+```
+
+Example:
+![In-Flight Requests](images/in_flight_requests.png)
\ No newline at end of file
diff --git a/prometheus/docs/images/avg_propagation_delay.png b/prometheus/docs/images/avg_propagation_delay.png
new file mode 100644
index 00000000..9e5b0748
Binary files /dev/null and b/prometheus/docs/images/avg_propagation_delay.png differ
diff --git a/prometheus/docs/images/batch_efficiency_ratio.png b/prometheus/docs/images/batch_efficiency_ratio.png
new file mode 100644
index 00000000..9c1ae296
Binary files /dev/null and b/prometheus/docs/images/batch_efficiency_ratio.png differ
diff --git a/prometheus/docs/images/empty_poll_responses.png b/prometheus/docs/images/empty_poll_responses.png
new file mode 100644
index 00000000..e34a7d80
Binary files /dev/null and b/prometheus/docs/images/empty_poll_responses.png differ
diff --git a/prometheus/docs/images/event_processing_duration.png b/prometheus/docs/images/event_processing_duration.png
new file mode 100644
index 00000000..efd93143
Binary files /dev/null and b/prometheus/docs/images/event_processing_duration.png differ
diff --git a/prometheus/docs/images/high_latency_event_processing.png b/prometheus/docs/images/high_latency_event_processing.png
new file mode 100644
index 00000000..4c3edf3c
Binary files /dev/null and b/prometheus/docs/images/high_latency_event_processing.png differ
diff --git a/prometheus/docs/images/in_flight_events.png b/prometheus/docs/images/in_flight_events.png
new file mode 100644
index 00000000..58e3a1bd
Binary files /dev/null and b/prometheus/docs/images/in_flight_events.png differ
diff --git a/prometheus/docs/images/in_flight_requests.png b/prometheus/docs/images/in_flight_requests.png
new file mode 100644
index 00000000..78342ebb
Binary files /dev/null and b/prometheus/docs/images/in_flight_requests.png differ
diff --git a/prometheus/docs/images/records_per_poll.png b/prometheus/docs/images/records_per_poll.png
new file mode 100644
index 00000000..f80a901c
Binary files /dev/null and b/prometheus/docs/images/records_per_poll.png differ
diff --git a/prometheus/docs/images/requests_processed.png b/prometheus/docs/images/requests_processed.png
new file mode 100644
index 00000000..8152bbe9
Binary files /dev/null and b/prometheus/docs/images/requests_processed.png differ
diff --git a/prometheus/docs/localstack_metrics.md b/prometheus/docs/localstack_metrics.md
new file mode 100644
index 00000000..96e74455
--- /dev/null
+++ b/prometheus/docs/localstack_metrics.md
@@ -0,0 +1,75 @@
+# LocalStack Metrics
+
+## LocalStack Core/Request Handling Metrics
+
+`localstack_request_processing_duration_seconds`
+
+- **Description:** Time spent processing LocalStack service requests. This is done at the handler chain and is calculated as the duration from first *request handler* to the final *response handler*.
+- **Labels:** `service`, `operation`, `status`, `status_code`
+- **Type:** histogram
+
+`localstack_in_flight_requests`
+
+- **Description:** Total number of currently in-flight requests. This is a live number, and will be influenced by the scraping interval.
+- **Labels:** `service`, `operation`
+- **Type:** gauge
+
+## LocalStack Event Poll Operation Metrics
+
+`localstack_records_per_poll`
+
+- **Description:** Number of records/events received in each poll operation
+- **Labels:** `event_source`, `event_target`
+- **Type:** histogram
+
+`localstack_poll_events_duration_seconds`
+
+- **Description:** Duration of each poll call in seconds
+- **Labels:** `event_source`, `event_target`
+- **Type:** histogram
+
+`localstack_poll_miss_total`
+
+- **Description:** Count of poll events with empty responses
+- **Labels:** `event_source`, `event_target`
+- **Type:** counter
+
+`localstack_batch_size_efficiency_ratio`
+
+- **Description:** Ratio of records received to configured maximum batch size
+- **Labels:** `event_source`, `event_target`
+- **Type:** histogram
+- **Note:** This is useful for finding whether the configured batch size is efficiently pulling records. A higher number indicates that a configured `BatchSize` could be increased.
+
+`localstack_batch_window_efficiency_ratio` (Not currently instrumented)
+
+- **Description:** Ratio of poll duration to configured maximum batch window length
+- **Labels:** `event_source`, `event_target`
+- **Type:** histogram
+- **Note:** Measures what proportion of the configured maximum batch window (set by `MaximumBatchingWindowInSeconds`) was actually used before returning. A lower ratio indicates that events were received quickly without needing to wait for the full window duration and that a window could be decreased.
+
+## LocalStack Event Processing Metrics
+
+`localstack_processed_events_total`
+
+- **Description:** Total number of events processed
+- **Labels:** `event_source`, `event_target`, `status`
+- **Type:** counter
+
+`localstack_in_flight_events`
+
+- **Description:** Total number of event batches currently being processed by the target
+- **Labels:** `event_source`, `event_target`
+- **Type:** gauge
+
+`localstack_event_propagation_delay_seconds`
+
+- **Description:** End-to-end latency between event creation and processing
+- **Labels:** `event_source`, `event_target`
+- **Type:** histogram
+
+`localstack_event_processing_errors_total`
+
+- **Description:** Total number of event processing errors
+- **Labels:** `event_source`, `event_target`, `error_type`
+- **Type:** counter
diff --git a/prometheus/docs/system_metrics.md b/prometheus/docs/system_metrics.md
new file mode 100644
index 00000000..97b8c810
--- /dev/null
+++ b/prometheus/docs/system_metrics.md
@@ -0,0 +1,67 @@
+# Sytem-level Metrics
+
+##  Garbage Collection Metrics
+
+`_gc_objects_collected_total`
+
+- **Description:** Number of objects collected during garbage collection
+- **Labels:** `generation`
+- **Type:** counter
+
+`_gc_objects_uncollectable_total`
+
+- **Description:** Number of uncollectable objects found during garbage collection
+- **Labels:** `generation`
+- **Type:** counter
+
+`_gc_collections_total`
+
+- **Description:** Number of times this generation was collected
+- **Labels:** `generation`
+- **Type:** counter
+
+##  Environment Metrics
+
+`_info`
+
+- **Description:**  platform information
+- **Labels:** `implementation`, `major`, `minor`, `patchlevel`, `version`
+- **Type:** gauge
+
+## Process Metrics
+
+`process_virtual_memory_bytes`
+
+- **Description:** Virtual memory size in bytes
+- **Labels:** none
+- **Type:** gauge
+
+`process_resident_memory_bytes`
+
+- **Description:** Resident memory size in bytes
+- **Labels:** none
+- **Type:** gauge
+
+`process_start_time_seconds`
+
+- **Description:** Start time of the process since unix epoch in seconds
+- **Labels:** none
+- **Type:** gauge
+
+`process_cpu_seconds_total`
+
+- **Description:** Total user and system CPU time spent in seconds
+- **Labels:** none
+- **Type:** counter
+
+`process_open_fds`
+
+- **Description:** Number of open file descriptors
+- **Labels:** none
+- **Type:** gauge
+
+`process_max_fds`
+
+- **Description:** Maximum number of open file descriptors
+- **Labels:** none
+- **Type:** gauge