Gavin McNair

@OleAkaNudelholz @oculus Try other USB3 ports. Mine only worked with the ports on the back of the machine. I thought my cables were bad.

A Fond Farewell to Oculus Go, Plus A New Distribution Path for Oculus Quest, Coming in 2021 // ocul.us/update

@oculus And the promised OTG support for the Go? (And the quest). Please at least let us use our defunct Go's for media.

@karlstoney @swade1987 @nickebbitt How many things are you scraping in the JMX exporter yaml config? Are you using one the the standard ones?

@gavinmcnair @swade1987 @nickebbitt might be able to comment more but we use it across our entire (massive) estate, including Kafka and Solr (Lucene), and some 300 apps as an agent

@karlstoney @swade1987 Not keen on using the agent since i dont want to affect the performance characteristics or reliability of the main Kafka. (and want to be able to measure them independently). Then again. Maybe its less resource intensive as a java agent

@swade1987 How are you measuring it’s resource usage as it’s just an agent in the main process? Curious as we use it too

@swade1987 @karlstoney If you use the httpserver version of the JMX exporter it gets all the stats directly from the Kafka process (via JMX) and just exposes them for prometheus to scrape

@SW_Railway the only thing thing reliable about you guys is that you are reliably running late every day. I can't remember the last time my Egham 7:12 train to Waterloo wasn't delayed or cancelled.

@swade1987 @joinmettle It's only very alpha and it's still going through changes but I'd appreciate feedback. I am also aware that storing both passwords and 2fa's together however secure defeats the purpose of 2fa in the first place.

@swade1987 @PrometheusIO @juliusvolz It looks like the right approach might be to extend github.com/prometheus/pro… then update github.com/ncabatoff/proc… to use it.

@swade1987 @PrometheusIO @juliusvolz It's probably fairly similar to measuring fd's open per process (versus nfiles for the number per process)

@swade1987 @PrometheusIO @juliusvolz It's an important value when you run anything storage which memory maps files. It needs adjusting for Kafka, Cassandra, elastic, etc...

@swade1987 @PrometheusIO @juliusvolz I think this can be obtained by counting the maps in /proc/<PID>/maps there I'd a single line or map in there. I've written code to get these values but because it's our process rather than node wide I'm not sure how it would be integrated.

@swade1987 @PrometheusIO @juliusvolz It might be slightly more complicated since max_map_count is the max count per process (which also is inherited into docker). What we really need is the count of the maps per process to see how close we are to max_map_count.

@bobacks I got one on launch day too. It is definitely heavy. Seriously heavy. Don't have the other problems though.

@StokkeLovers I know. We bought one but it is annoying. An Ikea high chair quarter of the price has them. It's an essential safety feature

@StokkeLovers My baby managed to fall out of our Tripp Trapp in the minute it took to get more food. Why do they not have straps?

@skyguideteam looks like I caught the wrong train. I'm going to be a bit late (just past Putney and have to change at clap ham junction)

@rnewstead Get well soon. What happened?

@rfletcherEW Happy birthday!