I recently got my node running and was checking through the dashboard to see about my performance. I noticed that my uptime on a few nodes were at 50%, which I simply assumed was due to some initial checks that failed during setup.
Sure enough, my uptime has continued to rise, but not nearly as rapidly as I would expect. Is there a way to find uptime logs of some capacity? Or rather, what all affects the uptime statistic? As far as I know, my machine should be maintaining 100% connectivity, and I haven’t had any network drops or restarts in the few days of running the node.
Where are you getting the uptime stats? I use uptimerobot.com which does a nice job tracking my nodes, up to 50 for free. It measures whether probed ports (tcp 28967 and up) are reachable by sending a packet every 5 minutes to each node.
If it’s a new node, or even an old node on a new satellite, I believe the uptime starts at 50% and eventually increases to 100%. It’s kind of a running average. Normal for a new node though.
I’m just using the local dashboard’s information for the node uptime. It makes sense that Storj itself should have the information for a node’s reputation/uptime available to us, instead of needing to use a third-party to monitor our nodes.
The problem with using an internal measurement is that too many externalities can affect your availability such as an ISP problem, your IP address changing and DDNS not being effective or a firewall issue. Uptimerobot helps measure that and alert you if there is a problem and you can find many posts in the forum about how people were able to realize this.
Now I suppose that the the Satellites could do this, but that would be sort of like the fox guarding the henhouse, I like to have a neutral third party validate my availability (although we do have to trust the satellites to properly audit us and determine our reputation. There are scripts you can run to see the vetting and audit results.
@stuberman
if you select a satellite you can see the uptime check % for that satellite
@brianlenz
getting to 100% uptime takes a long time… like a month of near flawless uptime… maybe two
the it will stay there and is pretty difficult to make it go down… unless if you have like very very long downtime… like days before it really shows… haven’t tried that tho…
It is curious how drastically different some of the satellites are, though. I actually have 100% uptime on 1 of the 6 satellites, but I’d assume that’s because I have next to no data for that one.
uptime checks sort of going merged with audits… so the more data you get from a satellite the faster you might see uptime checks go up… also some of the satellites mainly do test data or is being phased out… maybe… not sure… anyways the ones that do test data might be the only ones you get data from for the first month or so… until your node is vetted for being stable and able to deliver the stored data back without corruption and when requested.
thus you will most likely see that every time with a new node in the current version of the network, after a month or less you will be vetted and begin getting data from all satellites… i think… the vetting process is like 100 audit checks per satellite… but like i said, the test satellites may well go first… to make sure your node is good… before storing important stuff
