You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As the current telemetry only records basic success/error message from submodules, we should implement a more comprehensive recording system that will allow for much more debugging via the ground station.
Increasing the telemetry packet size introduces the following issues:
How will telemetry handle and store these new metrics
Use the current Log and Error classes to wrap around these metrics
Will the APRS be able to handle dumping all metrics at one time?
Described below
New Metric collection/reporting system
This new system includes 4 parts:
Telemetry
APRS
Core
Command Ingest
Instead of having dedicated Telemetry dumps every hour, there will be two separate processes regarding dumps:
Essential data will be emitted every 30 seconds
A large dump can be triggered through Command Ingest
Dataflow:
A new, metric.Metric class will be created to encapsulate data.
telemetry.enqueue will become more direct to deprecate telemetry.decide
a metric_stack will be added to telemetry
enqueue will directly push to these stacks
or call command_ingest.enqueue
core will call telemetry.emit, similar to telemetry.dump except only specific values from metric_stack are sent over to aprs
command_ingest will be able to call telemetry.dump, which will dump and clear all stacks through APRS
a new method telemetry.beacon will be called every 30 seconds by core and should read(not pop) the most recent, critical metrics from the metric_stack and send those through APRS.
The text was updated successfully, but these errors were encountered:
Description
As the current telemetry only records basic success/error message from submodules, we should implement a more comprehensive recording system that will allow for much more debugging via the ground station.
Increasing the telemetry packet size introduces the following issues:
Log
andError
classes to wrap around these metricsNew Metric collection/reporting system
This new system includes 4 parts:
Instead of having dedicated Telemetry dumps every hour, there will be two separate processes regarding dumps:
Dataflow:
metric.Metric
class will be created to encapsulate data.telemetry.enqueue
will become more direct to deprecatetelemetry.decide
metric_stack
will be added to telemetryenqueue
will directly push to these stackscommand_ingest.enqueue
core
will calltelemetry.emit
, similar totelemetry.dump
except only specific values frommetric_stack
are sent over toaprs
command_ingest
will be able to calltelemetry.dump
, which will dump and clear all stacks through APRStelemetry.beacon
will be called every 30 seconds bycore
and should read(not pop) the most recent, critical metrics from themetric_stack
and send those throughAPRS
.The text was updated successfully, but these errors were encountered: