Using message_interval on any message stops sys_status from being transmitted

[Crowdedlight]

Describe the bug
sys_status do not get transmitted after using a message_interval to limit any message. I have tried limiting different messages, but no change, and sys_status is still not being transmitted. I have also tried to send message_interval for sys_status to try and force it to 1 or 5hz. But didn't have any impact.

I checked with QGC and mavlink_inspector to verify if the message gets output.

To Reproduce
Steps to reproduce the behavior:

1. Drone switched on
2. Send a message_interval to limit some message
3. Open QGC and mavlink_inspector and sys_status isn't being transmitted, or with very low rate. (I sometimes got 1 maybe 2 messages over 30-40sec, but often didn't get any at all)

Rebooting the PX4 to output default messages makes the sys_status show up again, right until the next time you send message_interval.

Expected behavior
sys_status should not be "removed" when limiting some other message with message_interval.
Using message_interval for the sys_status message should correctly set the rate of the sys_status message.

Drone (please complete the following information):

• Hexacopter platform, PX4 1.9.2.

[dagar]

Could you try this in current master (or v1.10 beta)?

[Crowdedlight]

Tested on current master and tag: v1.10.0-beta1.
Same experience. After limiting any message I do not get sys_status. I sometimes get a single sys_status 15s after the limit, but nothing more than that.

We have also experienced that certain messages output by default can not be throttled by the message_interval command. Just gives MAV_RESULT_FAILED. Seems like it is because they are not defined as a streamlist_item if I read the source code correct.

We experienced it with the following msg ids:
MAVLINK_MSG_ID_RC_CHANNELS_SCALED = 34 # kept failing
MAVLINK_MSG_ID_RC_CHANNELS_RAW = 35 # kept failing
MAVLINK_MSG_ID_RC_CHANNELS_OVERRIDE = 70 # kept failing
MAVLINK_MSG_ID_BATTERY_STATUS = 147 # kept failing
MAVLINK_MSG_ID_ESTIMATOR_STATUS = 230 # kept failing

(I know this is probably another issue, but trying to give as much information as possible, if it helps)

[Crowdedlight]

Today we have experienced that using the message_interval mess with the output rate of global_position_int. When no message_interval has been sent so it output at default rates I receive global_position_int with 2hz, but as soon as message_interval is used, it drops down to inconsistent 0.3hz.

This has only been tested on PX4 1.9.2.

[dagar]

Could you try running mavlink status + mavlink status streams on the command line and capture the output before and after altering a message_interval?

[Crowdedlight]

The console didn't let me copy, so I captured it as images. It clearly show that something happens with sys_status which makes it go down to a rate of 0.05hz. Global_position_int also gets limited to 0.5hz without being told to. There also seems to be some odd values at system_time, altitude, extended_sys_state and estimator_status. However, they are all messages I try to limit to 1hz. I am gonna investigate if I send a wrong interval value for those messages.

Before limiting

Instance 0, is the one connected to my laptop when testing

After limiting

[Crowdedlight]

It would appear that if I send 1000000 as rate for altitude then it gets limited to 1hz.

I sent the following:

[[MAVLINK_MSG_ID_SYSTEM_TIME, 1000000],
[MAVLINK_MSG_ID_ALTITUDE, 1000000],
[MAVLINK_MSG_ID_EXTENDED_SYS_STATE, 2000000]]

Which resulted in the following output. So it seems like some of the streams use a different unit for rate? Or that it somehow is backwards?

EDIT: I am being stupid here. I somehow missed that the unit was in us, and not hz.... disregard that part of the error

[Crowdedlight]

It would appear that sys_status works as expected. The issue was me using the wrong unit for other messages. As such I assume it started to limits messages to stay below the overall limit of bytes/s, as it expected to send other messages with a million Hz.

Apologies for the stupid error. Gonna close the issue here. Thanks for the help :)