During the years, I encountered numerous lanterns in Qiryat Ata, that did a single or more on/off cycles, hot restriking at way lower temperatures than normal HPS in their state would restrike with external ignitors (Most restruck at LPS color), then returns to normal operation, and after several period of operation with such behavior, they stop working completely.

These "Temporary cyclers" in Qiryat Ata included:
1. http://www.lighting-gallery.net/gallery/displayimage.php?pos=-68471
2. http://www.lighting-gallery.net/gallery/displayimage.php?pos=-86702
3. http://www.lighting-gallery.net/gallery/displayimage.php?pos=-88018
4. http://www.lighting-gallery.net/gallery/displayimage.php?pos=-93888 (This lamp is on an anti-cycling ignitor, thats why it gives up after the second cycle)
5. http://www.lighting-gallery.net/gallery/displayimage.php?pos=-92669 (This one observed cycling non stop, but it cycles are similar to the rest [Hot restrike at LPS orange color, instead of at almost full brightness])
6. http://www.lighting-gallery.net/gallery/displayimage.php?pos=-99233

All of these cyclers begins their one or several cycles once the roadlighting turning on. Most of them extingushes before reaching to full output, and hot restrikes at way lower temperatures than they should be in their state with external ignitors. But once they stops cycling, they finally reaches full output, and remains operating steadily, until the next roadlighting turning on.

Today, when I photographed the new HPS lamps, that relamped the Koffers and the Aluras (The video will soon be uploaded), I observed the center lantern of the second triple lamppost, doing a single cycle in the style of all of the cyclers in the videos that I posted here. It was off for a long period of time, finally hot restruck at a pinkish white color, that changed to a dimmer than usual orange color, then return to normal operation.

What could causes HPS lamps to behave such?

All of these cyclers begins their one or several cycles once the roadlighting turning on. Most of them extingushes before reaching to full output, and hot restrikes at way lower temperatures than they should be in their state with external ignitors. But once they stops cycling, they finally reaches full output, and remains operating steadily, until the next roadlighting turning on.

Could'n it have something to do with an ambient temperature the lamp operates in? I mean when the lantern is still cold (at the moment of streetlights turned on), there's a 5 deg. C in it - for example (don't know current Qiryat Ata temperatures). But after several cycles, the lantern is already warm/hot - there could be 60 deg. C inside, so the lamp is cooled less...?

I don't think that this is related to the ambient temperature. I've seen such phenomenon also during the summer. Also, in Qiryat Atam the tempratures at the winter, not reaches below 10*C at the evening.

Nearing the EOL for HPS lamps, they will behave like this..cycling sometimes then working ok for a length of time, then finally giving in to constant cycling and death. Sometimes (but rarely) the orientation or position of the arc-tube might pose a problem in keeping the arc going.

But whenever I had a cycling fixture or lamp, it almost always seemed to be the photocontrol. It just happened to me yesterday when testing a newly refurbished MV fixture. The photocontrol turned off my fixture when I left the room after about 5 minutes (wanting to run the ballast & get both lamp & ballast hot). When I came back the lamp seemed dim and I realized it was getting brighter.. until it turned off..oh damned, I forgot to cover the photocontrol. Covering the sensor with black tape solved the cycling.

Have you see the videos I've attached? All of these HPS lamps, did one or several cycles, restruck at very low temperatures, and returned to normal operation without any cycling until the next evening.
Also, in Israel we haven't lanterns with PCs. The lanterns are group switched.