It appears to be under a DOS attack, or abusive usage
It is clearly mentioned here: https://www.stellar.org/developers/guides/concepts/test-net.html that Testnet should not be used for stress or load testing, but since a few days it stays stuck at 100% usage for long periods of time. It seems improbable that this is the result of "reasonable" usage, it's been running smoothly for months.
It would suck if reasonable users would have to run their own testnet for running minimalist unit tests. Something like an nginx reverse proxy with some rate limiting config could be all it takes to prevent such abuse https://www.nginx.com/blog/rate-limiting-nginx/
The transactions monopolizing testnet all look like the following:
RECENT OPERATIONS: TEST NETWORK API
Source Operation Details Time ago
GCKO payment 0.0000001 XLM » GBLH 4s
GCFG payment 0.0000001 XLM » GBLH 4s
GDX4 payment 0.0000001 XLM » GBLH 4s
GCN4 payment 0.0000001 XLM » GBLH 4s
GATA payment 0.0000001 XLM » GBLH 4s
GDZU payment 0.0000001 XLM » GBLH 4s
GDMS payment 0.0000001 XLM » GBLH 4s
GBNZ payment 0.0000001 XLM » GBLH 4s
GAZQ payment 0.0000001 XLM » GBLH 4s
GB5W payment 0.0000001 XLM » GBLH 4s
GDLZ payment 0.0000001 XLM » GBLH 4s
GARS payment 0.0000001 XLM » GBLH 4s
GAQF payment 0.0000001 XLM » GBLH 4s
GAH7 payment 0.0000001 XLM » GBLH 4s
GABR payment 0.0000001 XLM » GBLH 4s
GB24 payment 0.0000001 XLM » GBLH 4s
GBAB payment 0.0000001 XLM » GBLH 4s
GDUT payment 0.0000001 XLM » GBLH 4s
GDQX payment 0.0000001 XLM » GBLH 4s
GBDY payment 0.0000001 XLM » GBLH 4s
Technical issues, bugs, etc are usualy submited via github issues, not sure if that's the proper channel in this case, it's not a bug...
