Yet another issue, metrics.i2p was added to the 'I2P Community Sites' section, whereas the former (and future) stats.i2p was in the 'Network and Developer Information' section. It seems like the…
Yes - that needs to change. In the case of Metrics if I can get the auth from Phong, we can reuse the icon from the stats? In the case of StormyCloud - can we simply colourize the logo in the…
Also, both of the new sites - metrics and stormycloud - were added with the "plugin.png" icon which isn't a great match.
The requests should have each come in with a custom icon and associated…
@zzz thank you - Phong said he needs to dig it up, so I think we will be okay and if not, I will let you know with time to remove the entry so that it is not a block.
bump, shall we give up waiting and close this?
Removing blocker tag due to Metrics being removed from the homepage until we get the auth string.
Actually I screwed up the math and the history above. The ban throttler is 75 in 5 minutes or 15/minute. The history of the original drop throttler:
- The ORIGINAL drop throttler was 30 in 1…
Of course there's a huge difference between "not very often" and "not at all".
So a prerequisite for having shouldThrottle() return true always for subdbs is to absolutely positively prove it,…
I don't think subDb's should get lookups. There's still room for me to be wrong, but I think lookups should only go to the main DB.
- If you send a lookup to a subDb then what do you expect to…
After your change a2446e1d0c9c652be9adc1527738fc262988ada7 the throttlers will always be null in subdbs. But because of the way the throttle methods are:
boolean shouldThrottleLookup(H…
More correction of my OP about the "discriminate" comment and the "from" field:
Reference the I2NP spec, the from field is either the requesting router hash (if direct response requested), or the…
A simple example of how to consolidate back to one throttler (assuming your goal of a 7/minute rate and a burst of 10):
A single throttle of 14 in 2 minutes would be the same rate (7/minute)…
I have the addressbook-side-workaround implemented and tested.
However, I feel that this is hackish, and the correct fix is on the server side, to add the Apache directive to edit the sent etag…