Deployment

Network latency differences between cloud and on-premises environments affecting real-time data polling intervals is the most likely cause of these symptoms.

In a lift and shift migration, the application architecture remains unchanged, but the network topology changes significantly. When the monitoring system was on-premises, the sensors, dashboards, and database server were likely on the same local network with sub-millisecond latency. After migrating to the cloud, the physical distance between sensors (still on-premises) and the cloud-hosted application introduces network latency.

With polling intervals of 500 milliseconds, even small increases in network round-trip time can cause:
- Dashboard update delays (the 2-3 second delays reported)
- Missed sensor readings when poll requests time out or overlap
- Data gaps when the polling cycle cannot complete before the next cycle begins

This is a classic symptom of lift and shift migrations where real-time applications are moved to the cloud while data sources remain on-premises.

The other options are less likely because:

• The CPU allocation theory is weakened by the fact that the cloud infrastructure matches original server specifications exactly, and CPU bottlenecks would more likely cause consistent slowdowns rather than missing data.

• Database connection pooling issues would typically manifest as connection errors or timeouts during high-load periods, not consistent delays and missing readings during normal operation.

• Memory allocation differences between hypervisors could affect performance, but the symptoms described (delays and missing data at specific polling intervals) are more characteristic of network latency issues than memory-related caching problems.

The correct answer is to compare manifest file structures and segment naming conventions between legacy and new content to identify format inconsistencies introduced during migration.

Adaptive bitrate streaming (ABR) relies heavily on manifest files (such as HLS .m3u8 or DASH .mpd files) that contain references to different quality segments and tell the player which segments to request at various bitrates. The key clues in the scenario are:

1. New content works properly for adaptive streaming
2. Legacy content fails to switch between quality levels
3. Storage checksums verify file integrity (the actual video files are intact)
4. Encoding metadata appears identical

Since the video files themselves are verified as intact and encoding metadata matches, the issue likely lies in how the manifest files or segment references were migrated. During migration, manifest files may have been corrupted, paths may have changed, segment naming conventions may differ between environments, or URL structures may not have been properly updated. This would explain why playback works (the player can access content) but quality switching fails (the manifest cannot properly reference alternative bitrate segments).

The CDN configuration option is less likely because CDN cache hit rates are confirmed to meet expectations, indicating content is being served properly.

The transcoding pipeline logs option focuses on encoding parameters, but the question states encoding metadata appears identical between old and new content, making this less relevant.

The client-side player logs option examines whether the adaptive algorithm receives correct bandwidth measurements, but since new content works correctly with the same player and endpoints, the issue is not with bandwidth detection or client-side behavior - it is specific to legacy content structure.

The @batchSize(1) decorator in Bicep controls how many resources are deployed concurrently when using a for loop, and setting it to 1 achieves sequential deployment identical to ARM templates' copy element with serial mode.

The @sequential() decorator does not exist in Bicep syntax. The mode: 'serial' property approach incorrectly attempts to use ARM template JSON syntax within Bicep, which uses decorators instead. The @deploymentMode('sequential') decorator is also not a valid Bicep construct.

This question tests understanding of Bicep syntax for controlling resource deployment iteration and the migration path from ARM template copy elements to Bicep decorators.