Troubleshooting MSS endpoints in AWS Elemental MediaPackage

This could be due to the LookaheadCount setting. MSS holds back segments at the live edge until 2 future segments are available (this value is fixed and not configurable).

Solution: Ensure your player is requesting segments within the available window, accounting for the fixed 2-segment lookahead buffer.

Some older devices may have limitations with certain MSS features.

Solution: Try using the Full manifest layout instead of Compact, as it's more widely supported by legacy players.

MSS only supports PlayReady DRM.

Solution: Ensure your SPEKE key provider is correctly configured to provide PlayReady keys. For more information, see Content encryption and DRM in AWS Elemental MediaPackage.

The MSS manifest may not load if the endpoint is not properly configured.

Solution: Verify that you've selected the ISM container type when creating your endpoint. MSS manifests require the ISM container type.

This could be due to network issues or segment availability.

Solution: Check your network connection and ensure your CDN is properly configured for MSS content. Also, verify that your segment duration is appropriate for your content and network conditions.

Track the number of requests for manifests and segments to understand your audience size and viewing patterns.

Normal pattern: MSS clients typically request manifests less frequently than HLS clients, with a ratio of approximately 1 manifest request to 10-20 segment requests.

Warning signs: Sudden drops in request count may indicate playback issues or CDN problems.

Monitor the distribution of HTTP status codes to identify potential issues.

Normal pattern: Primarily 200 OK responses, with some 412 errors near the live edge due to the lookahead buffer (these are expected).

Warning signs: High rates of 5xx errors, unexpected 403 errors, or 404 errors for established content.

Track the time it takes to serve manifest and segment requests.

Normal pattern: Manifest requests typically have higher latency than segment requests due to dynamic generation.

Warning signs: Increasing latency trends or spikes above 500ms for segment delivery.

Monitor the volume of data being delivered to understand bandwidth usage and costs.

Normal pattern: Consistent patterns that align with your audience's viewing habits.

Warning signs: Unexpected spikes or drops that don't correlate with audience size changes.

Triggers when the percentage of 4xx or 5xx errors exceeds a threshold.

aws cloudwatch put-metric-alarm \
  --alarm-name MSS_HighErrorRate \
  --alarm-description "Alarm when MSS endpoint error rate exceeds 5%" \
  --metrics '[
    {
      "Id": "e1",
      "Expression": "(m2+m3)/m1*100",
      "Label": "Error Rate"
    },
    {
      "Id": "m1",
      "MetricStat": {
        "Metric": {
          "Namespace": "AWS/MediaPackage",
          "MetricName": "EgressRequestCount",
          "Dimensions": [
            { "Name": "Channel", "Value": "YourChannelName" },
            { "Name": "Origin", "Value": "YourOriginEndpointName" }
          ]
        },
        "Period": 300,
        "Stat": "Sum"
      },
      "ReturnData": false
    },
    {
      "Id": "m2",
      "MetricStat": {
        "Metric": {
          "Namespace": "AWS/MediaPackage",
          "MetricName": "EgressRequestCount",
          "Dimensions": [
            { "Name": "Channel", "Value": "YourChannelName" },
            { "Name": "Origin", "Value": "YourOriginEndpointName" },
            { "Name": "StatusCodeRange", "Value": "4xx" }
          ]
        },
        "Period": 300,
        "Stat": "Sum"
      },
      "ReturnData": false
    },
    {
      "Id": "m3",
      "MetricStat": {
        "Metric": {
          "Namespace": "AWS/MediaPackage",
          "MetricName": "EgressRequestCount",
          "Dimensions": [
            { "Name": "Channel", "Value": "YourChannelName" },
            { "Name": "Origin", "Value": "YourOriginEndpointName" },
            { "Name": "StatusCodeRange", "Value": "5xx" }
          ]
        },
        "Period": 300,
        "Stat": "Sum"
      },
      "ReturnData": false
    }
  ]' \
  --threshold 5 \
  --comparison-operator GreaterThanThreshold \
  --evaluation-periods 2 \
  --alarm-actions [your-sns-topic-arn]
                        

Triggers when the average latency exceeds a threshold.

aws cloudwatch put-metric-alarm \
  --alarm-name MSS_HighLatency \
  --alarm-description "Alarm when MSS endpoint latency exceeds 500ms" \
  --metric-name EgressResponseTime \
  --namespace AWS/MediaPackage \
  --statistic Average \
  --period 300 \
  --threshold 500 \
  --comparison-operator GreaterThanThreshold \
  --dimensions Name=Channel,Value=YourChannelName Name=Origin,Value=YourOriginEndpointName \
  --evaluation-periods 3 \
  --alarm-actions [your-sns-topic-arn]
                        

Triggers when request count drops significantly, which could indicate delivery issues.

aws cloudwatch put-metric-alarm \
  --alarm-name MSS_RequestDrop \
  --alarm-description "Alarm when MSS requests drop by more than 50%" \
  --metric-name EgressRequestCount \
  --namespace AWS/MediaPackage \
  --statistic Sum \
  --period 300 \
  --threshold 0 \
  --comparison-operator LessThanThreshold \
  --dimensions Name=Channel,Value=YourChannelName Name=Origin,Value=YourOriginEndpointName \
  --evaluation-periods 2 \
  --alarm-actions [your-sns-topic-arn] \
  --treat-missing-data breaching
                        

Possible causes:

Recommended action: Check input stream health, ensure segment duration works well with the fixed 2-fragment lookahead buffer, and review CDN cache settings.

Possible causes:

Recommended action: Review your CDN configuration, consider implementing or optimizing origin shield settings, and check for network bottlenecks.

Possible causes:

Recommended action: If these align with expected audience patterns, this is normal. If not, review your CDN TTL settings to ensure they're appropriate for your content type.

Occurs when the player requests a segment that cannot be found. This typically happens when the requested segment is not available or the URL is incorrect.

Common causes:

Occurs when lookahead fragments are not available for the requested segment. This indicates that MediaPackage does not have the lookahead fragments internally available.

Common causes:

Occurs when access to the requested resource is denied.

Common causes:

Indicates server-side issues with the MSS endpoint or underlying infrastructure.

Common causes: