I have built the circuit from this answer.
I have R1 = 10k and C1 = 6.8nF which seems to be giving me about 7.5kHz output.
The timer works just fine but from a cold startup, the output is wildly unstable for 2-3 seconds. My scope shows it oscillating at a much higher frequency than intended until it settles on the expected one.
Is this simply a quirk of the 555? This circuit is being testing in isolation with a known good power supply, and decoupling caps (although the behavior is identical with or without them). I have seen this behavior occur regardless of the input voltage. I could see a very brief period of instability but for me to notice for multiple seconds when looking at the output seems indicative of a design issue on my end.
Other thoughts are that the C1 I'm using is a generic ceramic, perhaps it's exhibiting some thermal issues, or the RC circuit is taking a long time to build enough charge to saturate the BJT?
I can easily design around this and just be robust to an initially unstable timer with downstream components, but I am curious about what might cause this and how I might solve it.