You've got me confused, while adding a monitor mode would be useful, it doesn't exist in the upstream lua project either, not sure where you got SQMA_TESTING from.
It is in my backlog though, together with a metrics exporter to dump stats into InfluxDB or any other tools where it'd be easy to ingest
You've got me confused, while adding a monitor mode would be useful, it doesn't exist in the upstream lua project either, not sure where you got `SQMA_TESTING` from.
It is in my backlog though, together with a metrics exporter to dump stats into InfluxDB or any other tools where it'd be easy to ingest
SQMA_TESTING ... wait what? Did it survive from the bash script days? Did I confuse myself?
I'll go do some digging, I was lacy and dumped a few old config backups at AI ... I will try to understand where it found it.
I do remember running some autorotate script in monitoring mode to create some nice Grafana graphs some 2 probably actually 3 years ago, let me find out what I was actually using.
Sorry for the "noise"
SQMA_TESTING ... wait what? Did it survive from the bash script days? Did I confuse myself?
I'll go do some digging, I was lacy and dumped a few old config backups at AI ... I will try to understand where it found it.
I do remember running some autorotate script in monitoring mode to create some nice Grafana graphs some 2 probably actually 3 years ago, let me find out what I was actually using.
Sorry for the "noise"
Apparently it's a side effect of me being lazy and asking AI to convert the cake-autorate config.sh over and then inquiring about monitor mode ... and boom we got a hallucination.
Sorry about that!
Ok that's totally on me!
Apparently it's a side effect of me being lazy and asking AI to convert the cake-autorate config.sh over and then inquiring about monitor mode ... and boom we got a hallucination.
Sorry about that!
I suspected it was something AI dreamt up yes ;-)
But it is something I wanted to add anyway, so it's not noise.
And yes I think it's the cake-autorate bash script you're thinking of, there you can set:
| adjust_dl_shaper_rate | enable (1) or disable (0) download shaping |
| adjust_ul_shaper_rate | enable (1) or disable (0) upload shaping |
to effectively put it in a monitor mode
What are you using for ingestion with Grafana? Like I said I have a side project to implement direct metrics/log export in sqm-autorate-rust itself. I toyed with InfluxDB because the line protocol there is very simple to work with, but I don't know what's the most popular these days
I suspected it was something AI dreamt up yes ;-)
But it is something I wanted to add anyway, so it's not noise.
And yes I think it's the cake-autorate bash script you're thinking of, there you can set:
| `adjust_dl_shaper_rate` | enable (1) or disable (0) download shaping |
| `adjust_ul_shaper_rate` | enable (1) or disable (0) upload shaping |
to effectively put it in a monitor mode
What are you using for ingestion with Grafana? Like I said I have a side project to implement direct metrics/log export in sqm-autorate-rust itself. I toyed with InfluxDB because the line protocol there is very simple to work with, but I don't know what's the most popular these days
Oh didn't see you figured it out yourself before responding, but yeah they do that a lot. People call them "professional lying machines" for a reason hehe
Oh didn't see you figured it out yourself before responding, but yeah they do that a lot. People call them "professional lying machines" for a reason hehe
my current setup uses InfluxDB and telegraf( docker with all relevant things mounted in that I want it to scrape) and Netdata as my still unfinished monitoring stack.
When I discovered occasional weird buffer bloat, I remembered the *-autorotate family and that I had some monitoring setup in the past while living with Starlink and decided to set up something like it again, to find out what's going on ... and that's how we got here.
What are you using for ingestion with Grafana?
Nothing( yet), 2/3 years ago according to my wiki I used: https://raw.githubusercontent.com/bairhys/prometheus-cake-autorate-exporter/main/prometheus-node-exporter-cake-autorate and [this](https://grafana.com/grafana/dashboards/18597-cake-autorate/) Grafana cake-autorate dashboard (ID 18597)
my current setup uses InfluxDB and telegraf( docker with all relevant things mounted in that I want it to scrape) and Netdata as my still unfinished monitoring stack.
When I discovered occasional weird buffer bloat, I remembered the *-autorotate family and that I had some monitoring setup in the past while living with Starlink and decided to set up something like it again, to find out what's going on ... and that's how we got here.
Aah got it.
Just added a monitor mode (dry run) and a metrics exporter
Binaries are at: https://build.lochnair.net/job/Bufferbloat/job/sqm-autorate-rust/job/master/
Note those dont have UCI enabled, since its not a OpenWrt build environment. I was testing using the env variables instead.
I updated the openwrt feed to use the new version, so you can build it with the Openwrt SDK if you wanna try
Just dumping the config definition here since I dont have example configs ready yet:
```rust
dry_run: Self::get::<bool>(
"SQMA_DRY_RUN",
"sqm-autorate.@advanced_settings[0].dry_run",
Some(false),
)?,
```
```rust
// Observability section
observability_enabled: Self::get::<bool>(
"SQMA_OBSERVABILITY_ENABLED",
"sqm-autorate.@observability[0].enabled",
Some(false),
)?,
observability_protocol: Self::get::<ObservabilityProtocol>(
"SQMA_OBSERVABILITY_PROTOCOL",
"sqm-autorate.@observability[0].protocol",
Some(ObservabilityProtocol::Udp),
)?,
observability_host: Self::get_optional::<String>(
"SQMA_OBSERVABILITY_HOST",
"sqm-autorate.@observability[0].host",
),
observability_port: Self::get::<u16>(
"SQMA_OBSERVABILITY_PORT",
"sqm-autorate.@observability[0].port",
Some(8089),
)?,
observability_batch_size: Self::get::<usize>(
"SQMA_OBSERVABILITY_BATCH_SIZE",
"sqm-autorate.@observability[0].batch_size",
Some(25),
)?,
observability_batch_timeout_ms: Self::get::<u64>(
"SQMA_OBSERVABILITY_BATCH_TIMEOUT_MS",
"sqm-autorate.@observability[0].batch_timeout_ms",
Some(100),
)?,
observability_export_ping_metrics: Self::get::<bool>(
"SQMA_OBSERVABILITY_EXPORT_PING_METRICS",
"sqm-autorate.@observability[0].export_ping_metrics",
Some(false),
)?,
observability_export_rate_metrics: Self::get::<bool>(
"SQMA_OBSERVABILITY_EXPORT_RATE_METRICS",
"sqm-autorate.@observability[0].export_rate_metrics",
Some(true),
)?,
observability_export_baseline_metrics: Self::get::<bool>(
"SQMA_OBSERVABILITY_EXPORT_BASELINE_METRICS",
"sqm-autorate.@observability[0].export_baseline_metrics",
Some(false),
)?,
observability_export_events: Self::get::<bool>(
"SQMA_OBSERVABILITY_EXPORT_EVENTS",
"sqm-autorate.@observability[0].export_events",
Some(true),
)?,
```
I used this when testing:
```
export SQMA_DOWNLOAD_INTERFACE=ifb4wan
export SQMA_UPLOAD_INTERFACE=wan
export SQMA_DOWNLOAD_BASE_KBITS=105000
export SQMA_UPLOAD_BASE_KBITS=150000
export SQMA_LOG_LEVEL=debug
export SQMA_SUPPRESS_STATISTICS=false
export SQMA_MEASUREMENT_TYPE=icmp-timestamps
export SQMA_NUM_REFLECTORS=5
export SQMA_TICK_INTERVAL=0.5
export SQMA_MIN_CHANGE_INTERVAL=0.5
export SQMA_DOWNLOAD_DELAY_MS=15
export SQMA_UPLOAD_DELAY_MS=15
export SQMA_HIGH_LOAD_LEVEL=0.8
export SQMA_REFLECTOR_LIST_FILE=/root/reflectors-icmp.csv
export SQMA_OBSERVABILITY_ENABLED=true
export SQMA_OBSERVABILITY_PROTOCOL=udp
export SQMA_OBSERVABILITY_HOST=192.168.1.2
export SQMA_OBSERVABILITY_PORT=8089
export SQMA_OBSERVABILITY_BATCH_SIZE=100
export SQMA_OBSERVABILITY_BATCH_TIMEOUT_MS=100
export SQMA_OBSERVABILITY_EXPORT_PING_METRICS=true
export SQMA_OBSERVABILITY_EXPORT_RATE_METRICS=true
export SQMA_OBSERVABILITY_EXPORT_BASELINE_METRICS=true
export SQMA_OBSERVABILITY_EXPORT_EVENTS=true
export SQMA_OBSERVABILITY_HOST_TAG=test-router
```
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
SQMA_TESTING=1
Seemed to have been ignored.
Ergo is it missing or am I using it wrong?
You've got me confused, while adding a monitor mode would be useful, it doesn't exist in the upstream lua project either, not sure where you got
SQMA_TESTINGfrom.It is in my backlog though, together with a metrics exporter to dump stats into InfluxDB or any other tools where it'd be easy to ingest
SQMA_TESTING ... wait what? Did it survive from the bash script days? Did I confuse myself?
I'll go do some digging, I was lacy and dumped a few old config backups at AI ... I will try to understand where it found it.
I do remember running some autorotate script in monitoring mode to create some nice Grafana graphs some 2 probably actually 3 years ago, let me find out what I was actually using.
Sorry for the "noise"
Ok that's totally on me!
Apparently it's a side effect of me being lazy and asking AI to convert the cake-autorate config.sh over and then inquiring about monitor mode ... and boom we got a hallucination.
Sorry about that!
I suspected it was something AI dreamt up yes ;-)
But it is something I wanted to add anyway, so it's not noise.
And yes I think it's the cake-autorate bash script you're thinking of, there you can set:
|
adjust_dl_shaper_rate| enable (1) or disable (0) download shaping ||
adjust_ul_shaper_rate| enable (1) or disable (0) upload shaping |to effectively put it in a monitor mode
What are you using for ingestion with Grafana? Like I said I have a side project to implement direct metrics/log export in sqm-autorate-rust itself. I toyed with InfluxDB because the line protocol there is very simple to work with, but I don't know what's the most popular these days
Oh didn't see you figured it out yourself before responding, but yeah they do that a lot. People call them "professional lying machines" for a reason hehe
What are you using for ingestion with Grafana?
Nothing( yet), 2/3 years ago according to my wiki I used: https://raw.githubusercontent.com/bairhys/prometheus-cake-autorate-exporter/main/prometheus-node-exporter-cake-autorate and this Grafana cake-autorate dashboard (ID 18597)
my current setup uses InfluxDB and telegraf( docker with all relevant things mounted in that I want it to scrape) and Netdata as my still unfinished monitoring stack.
When I discovered occasional weird buffer bloat, I remembered the *-autorotate family and that I had some monitoring setup in the past while living with Starlink and decided to set up something like it again, to find out what's going on ... and that's how we got here.
Aah got it.
Just added a monitor mode (dry run) and a metrics exporter
Binaries are at: https://build.lochnair.net/job/Bufferbloat/job/sqm-autorate-rust/job/master/
Note those dont have UCI enabled, since its not a OpenWrt build environment. I was testing using the env variables instead.
I updated the openwrt feed to use the new version, so you can build it with the Openwrt SDK if you wanna try
Just dumping the config definition here since I dont have example configs ready yet:
I used this when testing: