Hi @all,

I've the following output of a USR1 signal against the tor daemon in my log:

Oct 11 19:16:36.000 [notice] --------------- Dumping memory information:
Oct 11 19:16:36.000 [notice] In buffers for 2021 connections: 6103606 used/6770688 allocated
Oct 11 19:16:36.000 [notice]   For 2 OR listener connections: 0 used/0 allocated
Oct 11 19:16:36.000 [notice]   For 1269 OR connections: 478155 used/573440 allocated
Oct 11 19:16:36.000 [notice]   For 742 Exit connections: 5625383 used/6189056 allocated
Oct 11 19:16:36.000 [notice]   For 1 Socks listener connections: 0 used/0 allocated
Oct 11 19:16:36.000 [notice]   For 1 Directory listener connections: 0 used/0 allocated
Oct 11 19:16:36.000 [notice]   For 1 Directory connections: 68 used/8192 allocated
Oct 11 19:16:36.000 [notice]   For 4 CPU worker connections: 0 used/0 allocated
Oct 11 19:16:36.000 [notice]   For 1 Control listener connections: 0 used/0 allocated
Oct 11 19:16:36.000 [notice] In rephist: 567424 used by 3928 Tors.
Oct 11 19:16:36.000 [notice] In 7136 live descriptors: 9860168 bytes.  In 14728 old descriptors: 20229585 bytes.
Oct 11 19:16:36.000 [notice] 3317 cells allocated on 1203 circuits. 0 cells leaked.
Oct 11 19:16:36.000 [notice] Our DNS cache has 455 entries.
Oct 11 19:16:36.000 [notice] Our DNS cache size is approximately 184816 bytes.
Oct 11 19:16:36.000 [notice] mallinfo() said: arena=119304192, ordblks=6628, smblks=100, hblks=1, hblkhd=1052672, usmblks=0, fsmblks=6192, uordblks=89103552, fordblks=30200640, keepcost=4846928

For my monitoring of the relay I thought having a look at the built circuits would be the right approach and indicates a "healthy" relay. But then I stumbled over the "Connections" vs. "OR connections" vs "circuits" (I bolded the values).
Now I'm a little bit confused. What is the right approach to monitor the health of a (n exit) relay?

Thanks in advance,
~Josef